This model was trained by LiChenda using wsj0_2mix recipe in espnet. Python APISee https://github.com/espnet/espnet_model_zoo Evaluate in the recipegit clone https://github.com/espnet/espnet cd espnet git checkout 111c6cae7d9ce3f43dbbd2ba34ebaa0ca3efad10 pip install -e . cd egs2/wsj0_2mix/enh1 ./run.sh --skip_data_prep false --skip_train true --download_model LiChenda/wsj0_2mix_tasnet_8k_valid.si_snr.ave Results # RESULTS ## Environments - date: `Fri Aug 21 16:19:08 CST 2020` - python version: `3.7.6 (default, Jan 8 2020, 19:59:22) [GCC 7.3.0]` - espnet version: `espnet 0.7.0` - pytorch version: `pytorch 1.5.0` - Git hash: `caa78af2a309a5e169a1449d5d39ed0d0bdd9ef7` - Commit date: `Mon Aug 10 20:23:41 2020 +0800` ## tasnet_8k config: conf/tuning/train_enh_tasnet_lr01.yaml |dataset|PESQ|STOI|SAR|SDR|SIR| |---|---|---|---|---|---| |enhanced_cv_max_8k|3.38714|0.9557|17.7851|17.2729|28.2361| |enhanced_tt_max_8k|3.30601|0.953976|16.8024|16.1524|26.7848| ASR configconfig: conf/tuning/train_enh_tasnet_lr01.yaml print_config: false log_level: INFO dry_run: false iterator_type: chunk output_dir: exp/enh_train_enh_tasnet_lr01_raw ngpu: 1 seed: 0 num_workers: 4 num_att_plot: 3 dist_backend: nccl dist_init_method: env:// dist_world_size: 4 dist_rank: 1 local_rank: 1 dist_master_addr: localhost dist_master_port: 53979 dist_launcher: null multiprocessing_distributed: true cudnn_enabled: true cudnn_benchmark: false cudnn_deterministic: true collect_stats: false write_collected_feats: false max_epoch: 100 patience: 4 val_scheduler_criterion: - valid - loss early_stopping_criterion: - valid - loss - min best_model_criterion: - - valid - si_snr - max - - valid - loss - min keep_nbest_models: 1 grad_clip: 5.0 grad_noise: false accum_grad: 1 no_forward_run: false resume: true train_dtype: float32 log_interval: null pretrain_path: [] pretrain_key: [] num_iters_per_epoch: null batch_size: 16 valid_batch_size: null batch_bins: 1000000 valid_batch_bins: null train_shape_file: - exp/enh_stats_8k/train/speech_mix_shape - exp/enh_stats_8k/train/speech_ref1_shape - exp/enh_stats_8k/train/speech_ref2_shape valid_shape_file: - exp/enh_stats_8k/valid/speech_mix_shape - exp/enh_stats_8k/valid/speech_ref1_shape - exp/enh_stats_8k/valid/speech_ref2_shape batch_type: folded valid_batch_type: null fold_length: - 80000 - 80000 - 80000 sort_in_batch: descending sort_batch: descending multiple_iterator: false chunk_length: 32000 chunk_shift_ratio: 0.5 num_cache_chunks: 1024 train_data_path_and_name_and_type: - - dump/raw/tr_min_8k/wav.scp - speech_mix - sound - - dump/raw/tr_min_8k/spk1.scp - speech_ref1 - sound - - dump/raw/tr_min_8k/spk2.scp - speech_ref2 - sound valid_data_path_and_name_and_type: - - dump/raw/cv_min_8k/wav.scp - speech_mix - sound - - dump/raw/cv_min_8k/spk1.scp - speech_ref1 - sound - - dump/raw/cv_min_8k/spk2.scp - speech_ref2 - sound allow_variable_data_keys: false max_cache_size: 0.0 valid_max_cache_size: null optim: adam optim_conf: lr: 0.001 eps: 1.0e-08 weight_decay: 0 scheduler: reducelronplateau scheduler_conf: mode: min factor: 0.5 patience: 1 init: xavier_uniform model_conf: {} use_preprocessor: false enh: tasnet enh_conf: N: 256 L: 20 B: 256 H: 512 P: 3 X: 8 R: 4 num_spk: 2 norm_type: gLN causal: false mask_nonlinear: relu required: - output_dir distributed: true