--- tags: - espnet - audio - automatic-speech-recognition language: noinfo datasets: - librispeech_100 license: cc-by-4.0 --- ## ESPnet2 ASR model ### `jkang/espnet2_librispeech_100_conformer_char` This model was trained by jaekookang using librispeech_100 recipe in [espnet](https://github.com/espnet/espnet/). ### Demo: How to use in ESPnet2 ```bash cd espnet git checkout 82a0a0fa97b8a4a578f0a2c031ec49b3afec1504 pip install -e . cd egs2/librispeech_100/asr1 ./run.sh --skip_data_prep false --skip_train true --download_model jkang/espnet2_librispeech_100_conformer_char ``` # RESULTS ## Environments - date: `Thu Feb 24 17:47:04 KST 2022` - python version: `3.9.7 (default, Sep 16 2021, 13:09:58) [GCC 7.5.0]` - espnet version: `espnet 0.10.7a1` - pytorch version: `pytorch 1.10.1` - Git hash: `82a0a0fa97b8a4a578f0a2c031ec49b3afec1504` - Commit date: `Wed Feb 23 08:06:47 2022 +0900` ## asr_conformer_lr2e-3_warmup15k_amp_nondeterministic_char ### WER |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err| |---|---|---|---|---|---|---|---|---| |decode_asr_asr_model_valid.acc.ave/dev_clean|2703|54402|93.9|5.6|0.5|0.7|6.8|57.1| |decode_asr_asr_model_valid.acc.ave/dev_other|2864|50948|82.5|15.7|1.8|1.9|19.3|82.6| |decode_asr_asr_model_valid.acc.ave/test_clean|2620|52576|93.8|5.7|0.6|0.7|6.9|58.4| |decode_asr_asr_model_valid.acc.ave/test_other|2939|52343|82.2|15.9|2.0|1.7|19.5|83.6| ### CER |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err| |---|---|---|---|---|---|---|---|---| |decode_asr_asr_model_valid.acc.ave/dev_clean|2703|288456|98.3|1.0|0.7|0.7|2.4|57.1| |decode_asr_asr_model_valid.acc.ave/dev_other|2864|265951|93.3|4.1|2.6|1.9|8.7|82.6| |decode_asr_asr_model_valid.acc.ave/test_clean|2620|281530|98.3|1.0|0.7|0.6|2.3|58.4| |decode_asr_asr_model_valid.acc.ave/test_other|2939|272758|93.2|4.1|2.7|1.8|8.6|83.6| ### TER |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err| |---|---|---|---|---|---|---|---|---| ## ASR config
expand ``` config: conf/train_asr_char.yaml print_config: false log_level: INFO dry_run: false iterator_type: sequence output_dir: exp/asr_conformer_lr2e-3_warmup15k_amp_nondeterministic_char ngpu: 1 seed: 2022 num_workers: 4 num_att_plot: 0 dist_backend: nccl dist_init_method: env:// dist_world_size: null dist_rank: null local_rank: 0 dist_master_addr: null dist_master_port: null dist_launcher: null multiprocessing_distributed: false unused_parameters: false sharded_ddp: false cudnn_enabled: true cudnn_benchmark: false cudnn_deterministic: false collect_stats: false write_collected_feats: false max_epoch: 70 patience: null val_scheduler_criterion: - valid - loss early_stopping_criterion: - valid - loss - min best_model_criterion: - - valid - acc - max keep_nbest_models: 10 nbest_averaging_interval: 0 grad_clip: 5.0 grad_clip_type: 2.0 grad_noise: false accum_grad: 4 no_forward_run: false resume: true train_dtype: float32 use_amp: true log_interval: 400 use_matplotlib: true use_tensorboard: true use_wandb: false wandb_project: null wandb_id: null wandb_entity: null wandb_name: null wandb_model_log_interval: -1 detect_anomaly: false pretrain_path: null init_param: [] ignore_init_mismatch: false freeze_param: [] num_iters_per_epoch: null batch_size: 20 valid_batch_size: null batch_bins: 1600000 valid_batch_bins: null train_shape_file: - exp/asr_stats_raw_en_char_sp/train/speech_shape - exp/asr_stats_raw_en_char_sp/train/text_shape.char valid_shape_file: - exp/asr_stats_raw_en_char_sp/valid/speech_shape - exp/asr_stats_raw_en_char_sp/valid/text_shape.char batch_type: numel valid_batch_type: null fold_length: - 80000 - 150 sort_in_batch: descending sort_batch: descending multiple_iterator: false chunk_length: 500 chunk_shift_ratio: 0.5 num_cache_chunks: 1024 train_data_path_and_name_and_type: - - dump/raw/train_clean_100_sp/wav.scp - speech - kaldi_ark - - dump/raw/train_clean_100_sp/text - text - text valid_data_path_and_name_and_type: - - dump/raw/dev/wav.scp - speech - kaldi_ark - - dump/raw/dev/text - text - text allow_variable_data_keys: false max_cache_size: 0.0 max_cache_fd: 32 valid_max_cache_size: null optim: adam optim_conf: lr: 0.002 weight_decay: 1.0e-06 scheduler: warmuplr scheduler_conf: warmup_steps: 15000 token_list: - - - - E - T - A - O - N - I - H - S - R - D - L - U - M - C - W - F - G - Y - P - B - V - K - '''' - X - J - Q - Z - init: null input_size: null ctc_conf: dropout_rate: 0.0 ctc_type: builtin reduce: true ignore_nan_grad: true joint_net_conf: null model_conf: ctc_weight: 0.3 lsm_weight: 0.1 length_normalized_loss: false use_preprocessor: true token_type: char bpemodel: null non_linguistic_symbols: null cleaner: null g2p: null speech_volume_normalize: null rir_scp: null rir_apply_prob: 1.0 noise_scp: null noise_apply_prob: 1.0 noise_db_range: '13_15' frontend: default frontend_conf: n_fft: 512 win_length: 400 hop_length: 160 fs: 16k specaug: specaug specaug_conf: apply_time_warp: true time_warp_window: 5 time_warp_mode: bicubic apply_freq_mask: true freq_mask_width_range: - 0 - 27 num_freq_mask: 2 apply_time_mask: true time_mask_width_ratio_range: - 0.0 - 0.05 num_time_mask: 5 normalize: global_mvn normalize_conf: stats_file: exp/asr_stats_raw_en_char_sp/train/feats_stats.npz preencoder: null preencoder_conf: {} encoder: conformer encoder_conf: output_size: 256 attention_heads: 4 linear_units: 1024 num_blocks: 12 dropout_rate: 0.1 positional_dropout_rate: 0.1 attention_dropout_rate: 0.1 input_layer: conv2d normalize_before: true macaron_style: true rel_pos_type: latest pos_enc_layer_type: rel_pos selfattention_layer_type: rel_selfattn activation_type: swish use_cnn_module: true cnn_module_kernel: 31 postencoder: null postencoder_conf: {} decoder: transformer decoder_conf: attention_heads: 4 linear_units: 2048 num_blocks: 6 dropout_rate: 0.1 positional_dropout_rate: 0.1 self_attention_dropout_rate: 0.1 src_attention_dropout_rate: 0.1 required: - output_dir - token_list version: 0.10.7a1 distributed: false ```
### Citing ESPnet ```BibTex @inproceedings{watanabe2018espnet, author={Shinji Watanabe and Takaaki Hori and Shigeki Karita and Tomoki Hayashi and Jiro Nishitoba and Yuya Unno and Nelson Yalta and Jahn Heymann and Matthew Wiesner and Nanxin Chen and Adithya Renduchintala and Tsubasa Ochiai}, title={{ESPnet}: End-to-End Speech Processing Toolkit}, year={2018}, booktitle={Proceedings of Interspeech}, pages={2207--2211}, doi={10.21437/Interspeech.2018-1456}, url={http://dx.doi.org/10.21437/Interspeech.2018-1456} } ``` or arXiv: ```bibtex @misc{watanabe2018espnet, title={ESPnet: End-to-End Speech Processing Toolkit}, author={Shinji Watanabe and Takaaki Hori and Shigeki Karita and Tomoki Hayashi and Jiro Nishitoba and Yuya Unno and Nelson Yalta and Jahn Heymann and Matthew Wiesner and Nanxin Chen and Adithya Renduchintala and Tsubasa Ochiai}, year={2018}, eprint={1804.00015}, archivePrefix={arXiv}, primaryClass={cs.CL} } ```