You cannot select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

41 lines
1.2 KiB
Markdown

# Learning Nemo Toolkit
## Examples
I made a streaming example with `tts_stream.py`.
```
python tts_steam.py
```
Will take some time to load but eventually you will see
```
[NeMo I 2022-03-05 14:32:25 common:654] Instantiating model from pre-trained checkpoint
[NeMo I 2022-03-05 14:32:27 features:240] PADDING: 1
[NeMo I 2022-03-05 14:32:27 features:249] STFT using conv
[NeMo I 2022-03-05 14:32:27 features:251] STFT using exact pad
[NeMo I 2022-03-05 14:33:13 modelPT:376] Model SqueezeWaveModel was successfully restored from /home/toor/.cache/torch/NeMo/NeMo_1.0.0rc1/tts_squeezewave/d48f5835ac007ddb0c183bdbbbdace28/tts_squeezewave.nemo.
File tts_fifo_file already exists
Pipe text to tts_fifo_file
```
You can echo text steams to this file and the program will process them and play them on local speaker:
```
echo "I can't let you do that Dave." > tts_fifo_file
```
The first time it processes a text stream it will take a lot longer. It should get slightly faster after that first steam.
## Dependencies
I had
```
llvmlite-0.36.0-cp38-cp38-linux_aarch64.whl
onnxruntime_gpu-1.7.0-cp38-cp38-linux_aarch64.whl
torch-1.7.0-cp38-cp38-linux_aarch64.whl
```
In a dependencies folder. So I assume I hade these installed in then virtualenv manually