diff --git a/.compute b/.compute index 57e3805e..63dea7a7 100644 --- a/.compute +++ b/.compute @@ -10,7 +10,7 @@ wget https://www.dropbox.com/s/wqn5v3wkktw9lmo/install.sh?dl=0 -O install.sh sudo sh install.sh python3 setup.py develop # cp -R ${USER_DIR}/GermanData ../tmp/ -python3 distribute.py --config_path config_tacotron_de.json --data_path /data/rw/home/de_DE/ +python3 distribute.py --config_path config_libritts.json --data_path /data/rw/home/LibriTTS/train-clean-360/ # cp -R ${USER_DIR}/Mozilla_22050 ../tmp/ # python3 distribute.py --config_path config_tacotron_gst.json --data_path ../tmp/Mozilla_22050/ while true; do sleep 1000000; done diff --git a/config_libritts.json b/config_libritts.json index 84f48a75..10db4714 100644 --- a/config_libritts.json +++ b/config_libritts.json @@ -6,7 +6,7 @@ // Audio processing parameters "num_mels": 80, // size of the mel spec frame. "num_freq": 1025, // number of stft frequency levels. Size of the linear spectogram frame. - "sample_rate": 24000, // DATASET-RELATED: wav sample-rate. If different than the original data, it is resampled. + "sample_rate": 16000, // DATASET-RELATED: wav sample-rate. If different than the original data, it is resampled. "frame_length_ms": 50, // stft window length in ms. "frame_shift_ms": 12.5, // stft window hop-lengh in ms. "preemphasis": 0.98, // pre-emphasis to reduce spec noise and make it more structured. If 0.0, no -pre-emphasis. @@ -31,7 +31,7 @@ "reinit_layers": [], - "model": "Tacotron", // one of the model in models/ + "model": "Tacotron2", // one of the model in models/ "grad_clip": 1, // upper limit for gradients for clipping. "epochs": 1000, // total number of epochs to train. "lr": 0.0001, // Initial learning rate. If Noam decay is active, maximum learning rate. @@ -52,16 +52,16 @@ "separate_stopnet": true, // Train stopnet seperately if 'stopnet==true'. It prevents stopnet loss to influence the rest of the model. It causes a better model, but it trains SLOWER. "tb_model_param_stats": false, // true, plots param stats per layer on tensorboard. Might be memory consuming, but good for debugging. - "batch_size": 32, // Batch size for training. Lower values than 32 might cause hard to learn attention. + "batch_size": 16, // Batch size for training. Lower values than 32 might cause hard to learn attention. "eval_batch_size":16, - "r": 5, // Number of frames to predict for step. + "r": 1, // Number of frames to predict for step. "wd": 0.000001, // Weight decay weight. "checkpoint": true, // If true, it saves checkpoints per "save_step" "save_step": 1000, // Number of training steps expected to save traning stats and checkpoints. "print_step": 10, // Number of steps to log traning on console. "batch_group_size": 0, //Number of batches to shuffle after bucketing. - "run_eval": true, + "run_eval": false, "test_delay_epochs": 5, //Until attention is aligned, testing only wastes computation time. "test_sentences_file": null, // set a file to load sentences to be used for testing. If it is null then we use default english sentences. "data_path": "/home/erogol/Data/Libri-TTS/train-clean-360/", // DATASET-RELATED: can overwritten from command argument @@ -71,7 +71,7 @@ "min_seq_len": 6, // DATASET-RELATED: minimum text length to use in training "max_seq_len": 150, // DATASET-RELATED: maximum text length "output_path": "/media/erogol/data_ssd/Models/libri_tts/", // DATASET-RELATED: output path for all training outputs. - "num_loader_workers": 4, // number of training data loader processes. Don't set it too big. 4-8 are good values. + "num_loader_workers": 12, // number of training data loader processes. Don't set it too big. 4-8 are good values. "num_val_loader_workers": 4, // number of evaluation data loader processes. "phoneme_cache_path": "mozilla_us_phonemes", // phoneme computation is slow, therefore, it caches results in the given folder. "use_phonemes": true, // use phonemes instead of raw characters. It is suggested for better pronounciation.