Medix-AI/coqui-tts
Medix-AI
/
coqui-tts
mirror of https://github.com/coqui-ai/TTS.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

config update xz

This commit is contained in:
Eren Golge 2019-05-14 18:10:35 +02:00
parent b14c11572e
commit b94f71c56a
2 changed files with 6 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
config_cluster.json
View File

@ -1,6 +1,6 @@
{ {
"run_name": "mozilla-no-loc", "run_name": "mozilla-no-loc-fattn-stopnet",
"run_description": "using Bahdenau attention, with original prenet.", "run_description": "using forward attention, with original prenet, merged stopnet. Compare this with ",
"audio":{ "audio":{
// Audio processing parameters // Audio processing parameters
@ -42,12 +42,12 @@
"attention_norm": "softmax", // softmax or sigmoid. Suggested to use softmax for Tacotron2 and sigmoid for Tacotron. "attention_norm": "softmax", // softmax or sigmoid. Suggested to use softmax for Tacotron2 and sigmoid for Tacotron.
"prenet_type": "original", // ONLY TACOTRON2 - "original" or "bn". "prenet_type": "original", // ONLY TACOTRON2 - "original" or "bn".
"prenet_dropout": true, // ONLY TACOTRON2 - enable/disable dropout at prenet. "prenet_dropout": true, // ONLY TACOTRON2 - enable/disable dropout at prenet.
"use_forward_attn": false, // ONLY TACOTRON2 - if it uses forward attention. In general, it aligns faster. "use_forward_attn": true, // ONLY TACOTRON2 - if it uses forward attention. In general, it aligns faster.
"transition_agent": false, // ONLY TACOTRON2 - enable/disable transition agent of forward attention. "transition_agent": false, // ONLY TACOTRON2 - enable/disable transition agent of forward attention.
"location_attn": false, // ONLY TACOTRON2 - enable_disable location sensitive attention. It is enabled for TACOTRON by default. "location_attn": false, // ONLY TACOTRON2 - enable_disable location sensitive attention. It is enabled for TACOTRON by default.
"loss_masking": false, // enable / disable loss masking against the sequence padding. "loss_masking": false, // enable / disable loss masking against the sequence padding.
"enable_eos_bos_chars": false, // enable/disable beginning of sentence and end of sentence chars. "enable_eos_bos_chars": false, // enable/disable beginning of sentence and end of sentence chars.
"stopnet": false, // Train stopnet predicting the end of synthesis. "stopnet": true, // Train stopnet predicting the end of synthesis.
"separate_stopnet": false, // 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. "separate_stopnet": false, // 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.
"batch_size": 32, // Batch size for training. Lower values than 32 might cause hard to learn attention. "batch_size": 32, // Batch size for training. Lower values than 32 might cause hard to learn attention.

1
train.py
View File

@ -329,6 +329,7 @@ def evaluate(model, criterion, criterion_st, ap, current_step, epoch):
if num_gpus > 1: if num_gpus > 1:
postnet_loss = reduce_tensor(postnet_loss.data, num_gpus) postnet_loss = reduce_tensor(postnet_loss.data, num_gpus)
decoder_loss = reduce_tensor(decoder_loss.data, num_gpus) decoder_loss = reduce_tensor(decoder_loss.data, num_gpus)
if c.stopnet:
stop_loss = reduce_tensor(stop_loss.data, num_gpus) stop_loss = reduce_tensor(stop_loss.data, num_gpus)
avg_postnet_loss += float(postnet_loss.item()) avg_postnet_loss += float(postnet_loss.item())