Скачать или смотреть Configure Speaches for Captioning and Text to Speech (4 of 4)

In this video I'll take you through the steps in configuring the Speaches service to work with HERO1SHOT.
You can check out the HERO1SHOT videos here on this channel and find it in the Microsoft Store here: https://apps.microsoft.com/store/deta...

Links for Installation
Documentation for Speaches can be found here - commands are written for Linux but below and in the video I have converted to Windows 10/11.
https://speaches.ai/

===============================
Set up the Speaches Docker Container
===============================
Commands - Start by installing the Speaches Docker container by copying and then pasting into a Windows Command prompt:

docker run --detach --publish 8000:8000 --name speaches --volume hf-hub-cache:/home/ubuntu/.cache/huggingface/hub --gpus=all ghcr.io/speaches-ai/speaches:latest-cuda

If you're only going to caption audio/video files, stop here.
===============================================

==================
Using Text-to-Speech
==================
There are two sets of voices available - Kokoro and Piper. You can install one or both, it's your choice.

============
Kokoro voices
============

Each instruction line should be copied and pasted into a Windows Command prompt:
1. Open a Windows Command prompt.
2. Copy and paste the next line to set the variable:

set KOKORO_REVISION=c97b7bbc3e60f447383c79b2f94fee861ff156ac

3. Try this line next - if you get the error message, "No module named pip", go to step 4:

docker exec -it speaches huggingface-cli download hexgrad/Kokoro-82M --include "kokoro-v0_19.onnx" --revision %KOKORO_REVISION%

4. Run this to solve the "No module named pip" error - if you don't get it, skip to line 5:
docker exec -it speaches /home/ubuntu/speaches/.venv/bin/python3 -m ensurepip

5. Change YOURUSERNAME to your Windows username and if you get the error, "curl: (77) error setting certificate file: /etc/ssl/certs/ca-certificates.crt" then go to step six.

docker exec -it speaches curl --location --output C:\Users\YOURUSERNAME\.cache\huggingface\hub\models--hexgrad--Kokoro-82M\snapshots\%KOKORO_REVISION%\voices.bin https://github.com/thewh1teagle/kokor...

6. Run this to solve the "curl: (77) error" and change YOURUSERNAME to your Windows username. if you don't get this error, skip to line 7:

docker exec -it speaches curl -k --location --output C:\Users\YOURUSERNAME\.cache\huggingface\hub\models--hexgrad--Kokoro-82M\snapshots\$KOKORO_REVISION\voices.bin https://github.com/thewh1teagle/kokor...

7. Congratulations! You've successfully setup the Kokoro text-to-speech voices to work with HERO1SHOT.

============
Piper voices
============
Download all voices or a subset of voices. These commands are from https://speaches.ai/usage/text-to-spe... and have been updated to run in a Windows Command prompt.

Download/install all piper voices:

docker exec -it speaches huggingface-cli download rhasspy/piper-voices

Download/install a specific language (replace the EN for the language you want):

docker exec -it speaches huggingface-cli download rhasspy/piper-voices --include "en/**/*" "voices.json"

Download all qualities of a specific voice

docker exec -it speaches huggingface-cli download rhasspy/piper-voices --include "en/en_US/amy/**/*" "voices.json"

Download specific quality of a specific voice

docker exec -it speaches huggingface-cli download rhasspy/piper-voices --include "en/en_US/amy/medium/*" "voices.json"
00:00 Introduction
01:36 Marker 1