--- tags: - espnet - audio - automatic-speech-recognition language: it datasets: - voxforge license: cc-by-4.0 --- ## ESPnet2 ASR model ### `pyf98/voxforge_it_conformer_e15_linear1024` This model was trained by Yifan Peng using voxforge recipe in [espnet](https://github.com/espnet/espnet/). ### Demo: How to use in ESPnet2 Follow the [ESPnet installation instructions](https://espnet.github.io/espnet/installation.html) if you haven't done that already. ```bash cd espnet git checkout bf8c8f00194bdfed8ca388d8b20d14791b7d270e pip install -e . cd egs2/voxforge/asr1 ./run.sh --skip_data_prep false --skip_train true --download_model pyf98/voxforge_it_conformer_e15_linear1024 ``` # RESULTS ## Environments - date: `Thu Dec 29 01:59:25 EST 2022` - python version: `3.9.15 (main, Nov 24 2022, 14:31:59) [GCC 11.2.0]` - espnet version: `espnet 202211` - pytorch version: `pytorch 1.12.1` - Git hash: `bf8c8f00194bdfed8ca388d8b20d14791b7d270e` - Commit date: `Wed Dec 28 22:43:13 2022 -0500` ## asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse ### WER |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err| |---|---|---|---|---|---|---|---|---| |decode_asr_asr_model_valid.acc.ave/dt_it|1035|12587|70.2|24.6|5.2|3.3|33.1|94.7| |decode_asr_asr_model_valid.acc.ave/et_it|1103|13699|71.9|23.3|4.8|2.9|31.0|92.4| ### CER |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err| |---|---|---|---|---|---|---|---|---| |decode_asr_asr_model_valid.acc.ave/dt_it|1035|75494|92.9|3.9|3.2|1.8|9.0|94.7| |decode_asr_asr_model_valid.acc.ave/et_it|1103|81228|93.6|3.6|2.8|1.7|8.1|92.4| ### TER |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err| |---|---|---|---|---|---|---|---|---| ## ASR config
expand ``` config: conf/tuning/train_asr_conformer_e15_linear1024.yaml print_config: false log_level: INFO dry_run: false iterator_type: sequence output_dir: exp/asr_train_asr_conformer_e15_linear1024_raw_it_char_normalize_confnorm_varsFalse ngpu: 1 seed: 0 num_workers: 4 num_att_plot: 3 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: true collect_stats: false write_collected_feats: false max_epoch: 100 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: 1 no_forward_run: false resume: true train_dtype: float32 use_amp: true log_interval: null use_matplotlib: true use_tensorboard: true create_graph_in_tensorboard: false 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: 128 valid_batch_size: null batch_bins: 1000000 valid_batch_bins: null train_shape_file: - exp/asr_stats_raw_it_char/train/speech_shape - exp/asr_stats_raw_it_char/train/text_shape.char valid_shape_file: - exp/asr_stats_raw_it_char/valid/speech_shape - exp/asr_stats_raw_it_char/valid/text_shape.char batch_type: folded 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/tr_it/wav.scp - speech - sound - - dump/raw/tr_it/text - text - text valid_data_path_and_name_and_type: - - dump/raw/dt_it/wav.scp - speech - sound - - dump/raw/dt_it/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 scheduler: warmuplr scheduler_conf: warmup_steps: 10000 token_list: - - - - A - E - I - O - R - N - L - S - T - C - D - U - M - P - V - G - F - H - B - Q - Z - '''' - Ò - À - È - Ú - X - W - Í - É - Y - K - J - '1' - init: null input_size: null ctc_conf: dropout_rate: 0.0 ctc_type: builtin reduce: true ignore_nan_grad: null zero_infinity: true joint_net_conf: null 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' short_noise_thres: 0.5 frontend: default frontend_conf: fs: 16k specaug: null specaug_conf: {} normalize: global_mvn normalize_conf: stats_file: exp/asr_stats_raw_it_char/train/feats_stats.npz norm_vars: false model: espnet model_conf: ctc_weight: 0.3 lsm_weight: 0.1 length_normalized_loss: false preencoder: null preencoder_conf: {} encoder: conformer encoder_conf: output_size: 256 attention_heads: 4 linear_units: 1024 num_blocks: 15 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: 0.0 src_attention_dropout_rate: 0.0 preprocessor: default preprocessor_conf: {} required: - output_dir - token_list version: '202211' 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} } ```