maxmustermann/DeepLearningExamples
1
0
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity

Merge pull request #89 from GrzegorzKarchNV/readme-epochs

Browse source 
changed number of epochs in train scripts; removed number of epochs f…
This commit is contained in:
nvpstr 2019-07-01 11:44:01 +02:00 committed by GitHub
parent eabbedd1d4 7eb5eece14
commit f89dcca19d
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
3 changed files with 3 additions and 4 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

3
PyTorch/SpeechSynthesis/Tacotron2/README.md
View file

@ -137,8 +137,7 @@ Ensure your loss values are comparable to those listed in the table in the
Results section. For both models, the loss values are stored in the
`./output/nvlog.json` log file.
After you have trained the Tacotron 2 model for 1500 epochs and the
WaveGlow model for 800 epochs, you should get audio results similar to the
After you have trained the Tacotron 2 and WaveGlow models, you should get audio results similar to the
samples in the `./audio` folder. For details about generating audio, see the
[Inference process](#inference-process) section below.

2
PyTorch/SpeechSynthesis/Tacotron2/scripts/train_tacotron2.sh
View file

@ -1,2 +1,2 @@
mkdir -p output
python -m multiproc train.py -m Tacotron2 -o ./output/ -lr 1e-3 --epochs 2001 -bs 80 --weight-decay 1e-6 --grad-clip-thresh 1.0 --cudnn-benchmark=True --log-file ./output/nvlog.json --anneal-steps 500 1000 1500 --anneal-factor 0.1 --fp16-run
python -m multiproc train.py -m Tacotron2 -o ./output/ -lr 1e-3 --epochs 1500 -bs 80 --weight-decay 1e-6 --grad-clip-thresh 1.0 --cudnn-benchmark=True --log-file ./output/nvlog.json --anneal-steps 500 1000 1500 --anneal-factor 0.1 --fp16-run

2
PyTorch/SpeechSynthesis/Tacotron2/scripts/train_waveglow.sh
View file

@ -1,2 +1,2 @@
mkdir -p output
python -m multiproc train.py -m WaveGlow -o ./output/ -lr 1e-4 --epochs 2001 -bs 8 --segment-length 8000 --weight-decay 0 --grad-clip-thresh 65504.0 --epochs-per-checkpoint 50 --cudnn-benchmark=True --log-file ./output/nvlog.json --fp16-run
python -m multiproc train.py -m WaveGlow -o ./output/ -lr 1e-4 --epochs 1000 -bs 8 --segment-length 8000 --weight-decay 0 --grad-clip-thresh 65504.0 --epochs-per-checkpoint 50 --cudnn-benchmark=True --log-file ./output/nvlog.json --fp16-run