| ./tts.sh --lang gos --feats_type raw --audio_format flac --fs 22050 --n_fft 1024 --n_shift 256 --token_type char --cleaner none --g2p none --use_sid true --min_wav_duration 1 --max_wav_duration 20 --tts_task gan_tts --train_config conf/train_vits.yaml --train_set train_nodev --valid_set train_dev --test_sets train_dev --srctexts data/train/text --train_args '--use_wandb true --init_param downloads/espnet/kan-bayashi_ljspeech_vits/exp/tts_train_vits_raw_phn_tacotron_g2p_en_no_space/train.total_count.ave_10best.pth:tts:tts:tts.generator.text_encoder,tts.generator.posterior_encoder.input_conv' --tag vits --stage 5 "$@"; exit $? | |