--- tags: - espnet - audio - text-to-speech language: multilingual datasets: - masmultts license: cc-by-4.0 --- ## ESPnet2 TTS model ### `saefro991/tts_bytes_css10_7lang_textpretrain_residual_freeze` This model was trained by Takaaki-Saeki using masmultts 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 11a7d61312439111d4996d55935ede718d494262 pip install -e . cd egs2/masmultts/tts_byte_css10_adap_residual_freeze ./run.sh --skip_data_prep false --skip_train true --download_model saefro991/tts_bytes_css10_7lang_textpretrain_residual_freeze ``` ## TTS config
expand ``` config: conf/train.yaml print_config: false log_level: INFO dry_run: false iterator_type: sequence output_dir: exp/tts_train_raw_byte ngpu: 1 seed: 0 num_workers: 1 num_att_plot: 1 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: 200 patience: null val_scheduler_criterion: - valid - loss early_stopping_criterion: - valid - loss - min best_model_criterion: - - valid - loss - min - - train - loss - min keep_nbest_models: 3 nbest_averaging_interval: 0 grad_clip: 2.0 grad_clip_type: 2.0 grad_noise: false accum_grad: 4 no_forward_run: false resume: true train_dtype: float32 use_amp: false 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: - ../tts_pretrain_byte_residual/exp/tts_train_byte/2epoch.pth:tts_pretrain.encoder:tts.encoder - ../tts_pretrain_byte_residual/exp/tts_train_byte/2epoch.pth:tts_pretrain.lid_emb:tts.lid_emb ignore_init_mismatch: false freeze_param: - tts.encoder.adapter - tts.encoder.embed - tts.lid_emb num_iters_per_epoch: null batch_size: 20 valid_batch_size: null batch_bins: 400000 valid_batch_bins: null train_shape_file: - exp/tts_stats_raw_byte/train/text_shape.byte - exp/tts_stats_raw_byte/train/speech_shape valid_shape_file: - exp/tts_stats_raw_byte/valid/text_shape.byte - exp/tts_stats_raw_byte/valid/speech_shape batch_type: numel valid_batch_type: null fold_length: - 150 - 204800 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: - - /local/11399690.1.gpu/dump/raw/train/text - text - text - - /local/11399690.1.gpu/dump/raw/train/wav.scp - speech - sound - - /local/11399690.1.gpu/dump/xvector/train/xvector.scp - spembs - kaldi_ark - - /local/11399690.1.gpu/dump/raw/train/utt2lid - lids - text_int valid_data_path_and_name_and_type: - - /local/11399690.1.gpu/dump/raw/dev/text - text - text - - /local/11399690.1.gpu/dump/raw/dev/wav.scp - speech - sound - - /local/11399690.1.gpu/dump/xvector/dev/xvector.scp - spembs - kaldi_ark - - /local/11399690.1.gpu/dump/raw/dev/utt2lid - lids - text_int allow_variable_data_keys: false max_cache_size: 0.0 max_cache_fd: 32 valid_max_cache_size: null optim: adam optim_conf: lr: 1.0 scheduler: noamlr scheduler_conf: model_size: 512 warmup_steps: 50000 token_list: - - - '32' - '101' - '97' - '105' - '110' - '116' - '111' - '115' - '114' - '108' - '100' - '117' - '109' - '99' - '195' - '112' - '104' - '118' - '107' - '103' - '98' - '122' - '102' - '106' - '121' - '119' - '164' - '169' - '197' - '196' - '161' - '113' - '179' - '173' - '188' - '182' - '190' - '208' - '120' - '141' - '153' - '160' - '155' - '189' - '131' - '186' - '168' - '133' - '209' - '130' - '181' - '159' - '151' - '175' - '177' - '145' - '171' - '174' - '165' - '135' - '200' - '180' - '170' - '178' - '176' - '163' - '184' - '185' - '187' - '129' - '132' - '128' - '136' - '143' - '162' - '191' - '150' - '206' - '183' - '140' - '172' - '167' - '207' - '139' - '142' - '147' - '134' - '137' - '148' - '194' - '149' - '166' - '49' - '50' - '48' - '51' - '138' - '56' - '53' - '55' - '52' - '54' - '57' - '199' - '226' - '210' - '144' - '203' - '225' - '202' - '232' - '201' - '157' - '231' - '156' - '220' - odim: null model_conf: {} use_preprocessor: true token_type: byte bpemodel: null non_linguistic_symbols: null cleaner: null g2p: byte feats_extract: fbank feats_extract_conf: n_fft: 1024 hop_length: 256 win_length: null fs: 16000 fmin: 80 fmax: 7600 n_mels: 80 normalize: global_mvn normalize_conf: stats_file: exp/tts_stats_raw_byte/train/feats_stats.npz tts: transformer tts_conf: embed_dim: 0 eprenet_conv_layers: 0 eprenet_conv_filts: 0 eprenet_conv_chans: 0 dprenet_layers: 2 dprenet_units: 256 adim: 512 aheads: 8 elayers: 6 eunits: 1024 dlayers: 6 dunits: 1024 positionwise_layer_type: conv1d positionwise_conv_kernel_size: 1 postnet_layers: 5 postnet_filts: 5 postnet_chans: 256 spk_embed_dim: 192 spk_embed_integration_type: add use_gst: true gst_heads: 4 gst_tokens: 16 use_masking: true bce_pos_weight: 5.0 use_scaled_pos_enc: true encoder_normalize_before: true decoder_normalize_before: true reduction_factor: 1 init_type: xavier_uniform init_enc_alpha: 1.0 init_dec_alpha: 1.0 eprenet_dropout_rate: 0.0 dprenet_dropout_rate: 0.5 postnet_dropout_rate: 0.5 transformer_enc_dropout_rate: 0.1 transformer_enc_positional_dropout_rate: 0.1 transformer_enc_attn_dropout_rate: 0.1 transformer_dec_dropout_rate: 0.1 transformer_dec_positional_dropout_rate: 0.1 transformer_dec_attn_dropout_rate: 0.1 transformer_enc_dec_attn_dropout_rate: 0.1 use_guided_attn_loss: true num_heads_applied_guided_attn: 2 num_layers_applied_guided_attn: 2 modules_applied_guided_attn: - encoder-decoder guided_attn_loss_sigma: 0.4 guided_attn_loss_lambda: 10.0 langs: 21 lang_family_encoding: false num_lang_family: 7 use_adapter: true adapter_type: residual use_encoder_w_lid: true pitch_extract: null pitch_extract_conf: {} pitch_normalize: null pitch_normalize_conf: {} energy_extract: null energy_extract_conf: {} energy_normalize: null energy_normalize_conf: {} required: - output_dir - token_list version: '202209' 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} } @inproceedings{hayashi2020espnet, title={{Espnet-TTS}: Unified, reproducible, and integratable open source end-to-end text-to-speech toolkit}, author={Hayashi, Tomoki and Yamamoto, Ryuichi and Inoue, Katsuki and Yoshimura, Takenori and Watanabe, Shinji and Toda, Tomoki and Takeda, Kazuya and Zhang, Yu and Tan, Xu}, booktitle={Proceedings of IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)}, pages={7654--7658}, year={2020}, organization={IEEE} } ``` 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} } ```