inference_audio_tagging/log-decode-epoch-30-avg-6-use-averaged-model-2023-11-30-12-15-21

2023-11-30 12:15:21,528 INFO [inference_audio_tagging.py:316] Evaluation started
2023-11-30 12:15:21,529 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': '2b2ac14b326d61d79d04e53fbd69b1ff6d630411', 'k2-git-date': 'Thu Aug 24 05:58:26 2023', 'lhotse-version': '1.16.0', 'torch-version': '2.0.1+cu117', 'torch-cuda-available': True, 'torch-cuda-version': '11.7', 'python-version': '3.1', 'icefall-git-branch': 'multi_KD', 'icefall-git-sha1': '02ab7e51-dirty', 'icefall-git-date': 'Tue Nov 28 15:55:57 2023', 'icefall-path': '/star-data/xiaoyu/icefall_multi_KD', 'k2-path': '/star-xy/softwares/k2_development/k2/k2/python/k2/__init__.py', 'lhotse-path': '/star-xy/softwares/anaconda3/envs/multi_KD/lib/python3.10/site-packages/lhotse/__init__.py', 'hostname': 'de-74279-k2-train-1-1220091118-57c4d55446-mvd6x', 'IP address': '10.177.22.19'}, 'epoch': 30, 'iter': 0, 'avg': 6, 'use_averaged_model': True, 'exp_dir': PosixPath('multi_KD/exp_train_asr_full_libri1_do_audio_tagging1_as_unbalanced_scale1.0'), 'trained_with_distillation': False, 'trained_with_multitask': True, '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': False, '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': '16,32,64,-1', 'left_context_frames': '64,128,256,-1', 'use_transducer': True, 'use_ctc': False, 'speaker_input_idx': 2, 'whisper_dim': 768, 'num_codebooks': 32, 'mvq_kd_layer_idx': -1, 'use_subsampled_output': True, 'full_libri': True, 'mini_libri': False, 'use_voxceleb': False, 'voxceleb_subset': 'vox1', 'use_libriheavy': False, 'libriheavy_subset': 'small', 'use_audioset': False, 'audioset_subset': 'balanced', 'manifest_dir': PosixPath('data/fbank'), 'max_duration': 500, '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': False, 'whisper_mvq': False, 'beats_ckpt': 'data/models/BEATs/BEATs_iter3_plus_AS2M_finetuned_on_AS2M_cpt2.pt', 'whisper_version': 'small.en', '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_train_asr_full_libri1_do_audio_tagging1_as_unbalanced_scale1.0/inference_audio_tagging'), 'suffix': 'epoch-30-avg-6-use-averaged-model'}
2023-11-30 12:15:21,529 INFO [inference_audio_tagging.py:324] About to create model
2023-11-30 12:15:22,147 INFO [inference_audio_tagging.py:402] Calculating the averaged model over epoch range from 24 (excluded) to 30
2023-11-30 12:15:28,043 INFO [inference_audio_tagging.py:420] Number of model parameters: 65819362
2023-11-30 12:15:28,043 INFO [kd_datamodule.py:840] About to get the audioset eval cuts.
2023-11-30 12:15:28,051 INFO [kd_datamodule.py:534] About to create dev dataset
2023-11-30 12:15:28,617 INFO [kd_datamodule.py:555] About to create dev dataloader
2023-11-30 12:15:32,365 INFO [inference_audio_tagging.py:289] Processed 100 cuts already.
2023-11-30 12:15:36,137 INFO [zipformer.py:1877] name=None, attn_weights_entropy = tensor([2.1745, 1.8189, 3.4013, 2.8265, 3.6835, 3.6652, 3.1386, 3.1459],
       device='cuda:0')
2023-11-30 12:15:36,400 INFO [zipformer.py:1877] name=None, attn_weights_entropy = tensor([4.9719, 3.7998, 4.9852, 4.3820], device='cuda:0')
2023-11-30 12:15:36,831 INFO [inference_audio_tagging.py:289] Processed 1100 cuts already.
2023-11-30 12:15:41,319 INFO [inference_audio_tagging.py:289] Processed 2101 cuts already.
2023-11-30 12:15:43,345 INFO [zipformer.py:1877] name=None, attn_weights_entropy = tensor([2.8943, 3.9694, 3.7684, 3.1279], device='cuda:0')
2023-11-30 12:15:45,400 INFO [inference_audio_tagging.py:289] Processed 3101 cuts already.
2023-11-30 12:15:46,204 INFO [zipformer.py:1877] name=None, attn_weights_entropy = tensor([5.1761, 2.4422, 5.0019, 2.8404], device='cuda:0')
2023-11-30 12:15:46,561 INFO [zipformer.py:1877] name=None, attn_weights_entropy = tensor([1.5234, 3.0638, 2.7479, 2.7318, 3.3604, 3.4139, 3.0643, 3.6216],
       device='cuda:0')
2023-11-30 12:15:49,875 INFO [inference_audio_tagging.py:289] Processed 4101 cuts already.
2023-11-30 12:15:52,131 INFO [zipformer.py:1877] name=None, attn_weights_entropy = tensor([4.9854, 3.7455, 4.9535, 4.5031], device='cuda:0')
2023-11-30 12:15:53,104 INFO [zipformer.py:1877] name=None, attn_weights_entropy = tensor([4.2608, 4.2369, 4.4779, 4.4558], device='cuda:0')
2023-11-30 12:15:54,014 INFO [inference_audio_tagging.py:289] Processed 5101 cuts already.
2023-11-30 12:15:58,054 INFO [inference_audio_tagging.py:289] Processed 6101 cuts already.
2023-11-30 12:16:02,321 INFO [inference_audio_tagging.py:289] Processed 7101 cuts already.
2023-11-30 12:16:02,453 INFO [zipformer.py:1877] name=None, attn_weights_entropy = tensor([3.8484, 1.3301, 3.6270, 3.0730, 2.8808, 3.1913, 3.1312, 3.2145],
       device='cuda:0')
2023-11-30 12:16:06,410 INFO [inference_audio_tagging.py:289] Processed 8101 cuts already.
2023-11-30 12:16:07,301 INFO [zipformer.py:1877] name=None, attn_weights_entropy = tensor([4.6009, 3.6360, 3.9462, 3.4832], device='cuda:0')
2023-11-30 12:16:10,440 INFO [inference_audio_tagging.py:289] Processed 9101 cuts already.
2023-11-30 12:16:13,772 INFO [zipformer.py:1877] name=None, attn_weights_entropy = tensor([3.8172, 1.3767, 3.6141, 3.0482, 2.8630, 3.1663, 3.0868, 3.1847],
       device='cuda:0')
2023-11-30 12:16:14,464 INFO [inference_audio_tagging.py:289] Processed 10101 cuts already.
2023-11-30 12:16:16,211 INFO [zipformer.py:1877] name=None, attn_weights_entropy = tensor([3.9844, 3.1880, 2.9109, 3.2204, 3.4031, 2.8323, 3.3993, 2.6284],
       device='cuda:0')
2023-11-30 12:16:17,435 INFO [zipformer.py:1877] name=None, attn_weights_entropy = tensor([4.4718, 3.1547, 3.6060, 3.4538], device='cuda:0')
2023-11-30 12:16:18,555 INFO [inference_audio_tagging.py:289] Processed 11101 cuts already.
2023-11-30 12:16:21,267 INFO [zipformer.py:1877] name=None, attn_weights_entropy = tensor([2.9200, 4.0222, 3.7380, 3.1246], device='cuda:0')
2023-11-30 12:16:23,137 INFO [inference_audio_tagging.py:289] Processed 12101 cuts already.
2023-11-30 12:16:23,678 INFO [zipformer.py:1877] name=None, attn_weights_entropy = tensor([2.1587, 3.0726, 3.2987, 2.9865, 3.7040, 3.7674, 3.2999, 3.1897],
       device='cuda:0')
2023-11-30 12:16:27,414 INFO [inference_audio_tagging.py:289] Processed 13101 cuts already.
2023-11-30 12:16:31,568 INFO [inference_audio_tagging.py:289] Processed 14101 cuts already.
2023-11-30 12:16:36,085 INFO [inference_audio_tagging.py:289] Processed 15101 cuts already.
2023-11-30 12:16:36,588 INFO [inference_audio_tagging.py:290] Finish collecting audio logits
2023-11-30 12:16:40,203 INFO [inference_audio_tagging.py:453] mAP for audioset eval is: 0.41664740055105315
2023-11-30 12:16:40,203 INFO [inference_audio_tagging.py:455] Done