Published August 1, 2021
| Version v1
Other
Open
ESPnet2 pretrained model, kan-bayashi/hui_acg_tts_train_tacotron2_raw_hokuspokus_phn_espeak_ng_german_train.loss.ave, fs=24000, lang=de
Creators
Description
This model was trained by kan-bayashi using hui_acg/tts1 recipe in espnet.
- Python API
See https://github.com/espnet/espnet_model_zoo
- Evaluate in the recipe
git clone https://github.com/espnet/espnet cd espnet git checkout dd737de893711a2b918bb95f94887d38b9353d48 pip install -e . cd egs2/hui_acg/tts1 # Download the model file here ./run.sh --skip_data_prep false --skip_train true --download_model kan-bayashi/hui_acg_tts_train_tacotron2_raw_hokuspokus_phn_espeak_ng_german_train.loss.ave
- Config
config: conf/tuning/train_tacotron2.yaml print_config: false log_level: INFO dry_run: false iterator_type: sequence output_dir: exp/tts_train_tacotron2_raw_phn_none ngpu: 1 seed: 0 num_workers: 1 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: 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: 5 grad_clip: 1.0 grad_clip_type: 2.0 grad_noise: false accum_grad: 1 no_forward_run: false resume: true train_dtype: float32 use_amp: false log_interval: null 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: 500 batch_size: 20 valid_batch_size: null batch_bins: 3750000 valid_batch_bins: null train_shape_file: - exp/tts_stats_raw_phn_none/train/text_shape.phn - exp/tts_stats_raw_phn_none/train/speech_shape valid_shape_file: - exp/tts_stats_raw_phn_none/valid/text_shape.phn - exp/tts_stats_raw_phn_none/valid/speech_shape batch_type: numel valid_batch_type: null fold_length: - 150 - 240000 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/hokuspokus_tr_no_dev_phn/text - text - text - - dump/raw/hokuspokus_tr_no_dev_phn/wav.scp - speech - sound valid_data_path_and_name_and_type: - - dump/raw/hokuspokus_dev_phn/text - text - text - - dump/raw/hokuspokus_dev_phn/wav.scp - speech - sound allow_variable_data_keys: false max_cache_size: 0.0 max_cache_fd: 32 valid_max_cache_size: null optim: adam optim_conf: lr: 0.001 eps: 1.0e-06 weight_decay: 0.0 scheduler: null scheduler_conf: {} token_list: - - - n - ə - t - d - ɾ - l - s - m - f - z - ɪ - ɡ - ',' - v - ɛ - ç - b - k - ɜ - ˈa - iː - r - ʃ - ts - h - ˈɛ - ˈɪ - a - ʊ - ˈaɪ - . - p - ˈiː - ˈeː - ˌaɪ - ˈɑː - x - eː - ɑː - ŋ - ˈɔ - ˈaʊ - uː - ˈoː - ˈʊ - ɔ - aɪ - ˌa - ˈuː - ˌiː - ˌɪ - ˌɛ - oː - ˌɑː - ˈy - j - ˈɔø - aʊ - ˌɔ - ˌaʊ - ˈyː - ˌoː - ˈɛː - ˌeː - ˌʊ - '!' - ˈœ - ˈɑ - ˈøː - ˌyː - '?' - ':' - ˈ?? - yː - ?? - ˌuː - ɔø - ˌɑ - ˌ?? - ˌy - ˌɛː - ˌɔø - pf - y - ˌøː - ˌœ - ɛː - øː - œ - dʒ - i - ʒ - '1' - ˈə - (en) - (de) - ˌə - tʃ - ˈiə - odim: null model_conf: {} use_preprocessor: true token_type: phn bpemodel: null non_linguistic_symbols: null cleaner: null g2p: espeak_ng_german feats_extract: fbank feats_extract_conf: fs: 24000 fmin: 80 fmax: 7600 n_mels: 80 hop_length: 300 n_fft: 2048 win_length: 1200 normalize: global_mvn normalize_conf: stats_file: exp/tts_stats_raw_phn_none/train/feats_stats.npz tts: tacotron2 tts_conf: embed_dim: 512 elayers: 1 eunits: 512 econv_layers: 3 econv_chans: 512 econv_filts: 5 atype: location adim: 512 aconv_chans: 32 aconv_filts: 15 cumulate_att_w: true dlayers: 2 dunits: 1024 prenet_layers: 2 prenet_units: 256 postnet_layers: 5 postnet_chans: 512 postnet_filts: 5 output_activation: null use_batch_norm: true use_concate: true use_residual: false dropout_rate: 0.5 zoneout_rate: 0.1 reduction_factor: 1 spk_embed_dim: null use_masking: true bce_pos_weight: 5.0 use_guided_attn_loss: true guided_attn_loss_sigma: 0.4 guided_attn_loss_lambda: 1.0 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: 0.10.0 distributed: false
Files
tts_train_tacotron2_raw_hokuspokus_phn_espeak_ng_german_train.loss.ave.zip
Files
(107.6 MB)
Name | Size | Download all |
---|---|---|
md5:906605cf1a6496b6c59c95deab13b89e
|
107.6 MB | Preview Download |
Additional details
Related works
- Is supplement to
- https://github.com/espnet/espnet (URL)