AllTalk v1.7b

Updated timestamp to uuid on narrator/character file to avoid a possible race condition on file naming.
Added link in built in documentation to the "AllTalk Minor updates/bug fixes list"

AllTalk v1.7a

Manged to drop the Pydantic version down to 1.10.3

AllTalk_v1.7

Added the new API suite
Added model Finetuning
Updated all documentation
A couple of minor file updates.
Small bug fix on file generation in instruct mode with no character name available (thanks to @danielwburch)

If you download this file and extract it to your extensions folder. You have to remove the - from the extracted folder name. Github auto names the files with a - in them and I cannot change that. Best just to rename the extracted folder to alltalk_tts

AllTalk v1.5d

Built in documentation cleaned up and re-written where necessary.
Added a basic diagnostic utility diagnostic.py that creates screen output and diagnostics.log
Improved the error message for slow start-up of the subprocess.
Minor changes to documentation.
Small bug fix with IP/port not updating when custom address used.
Changed config.json to confignew.json due to having to re-name a variable as the new version of pydantic was calming there could be a clash, so changed the variable to avoid any possible conflict now or in future.
Corrected issue in .gitignore file.
DeepSpeed v11.x support with Windows.
Voice notifications on change of model settings.
Improved intonation and pronunciation.
Demo area within the settings.
Added additional API's (to be documented).

If you download this file and extract it to your extensions folder. You have to remove the - from the extracted folder name. Github auto names the files with a - in them and I cannot change that. Best just to rename the extracted folder to alltalk_tts

DeepSpeed v11.2 wheel files

Using the incorrect version for your system will cause AllTalk to have start up issues, so you may wish to check AllTalk is starting ok before installing DeepSpeed.

Back to the DeepSpeed install instructions

THESE ARE COMPILED FOR PyTorch 2.1.x and not the new 2.2.x

Python 3.11.x
DeepSpeed v11.2 for CUDA 11.8 and Python 3.11.x
DeepSpeed v11.2 for CUDA 12.1 and Python 3.11.x

Python 3.10.x (for a custom Python environment)
DeepSpeed v11.2 for CUDA 11.8 and Python 3.10.x
DeepSpeed v11.2 for CUDA 12.1 and Python 3.10.x

You can check the version you are using by running cmd_windows.bat in the Text-generation-webui folder and then:

python --version to get the Python version

pip show torch to get the CUDA version.

NOTE: You DO NOT need to set Text-generation-webUI's --deepspeed setting for AllTalk to be able to use DeepSpeed. These are two completely separate things and incorrectly setting that on Text-generation-webUI may cause other complications.

AllTalk v1.4

Has a rebuilt narration/character code block, along with better text filtering. This results in a much cleaner output with the end speech generated as well as a bit better control over the narrator function and what to do when the AI delivers text that isn't within double quote or asterisks.
(a couple of minor updates, mostly documentation)

AllTalk (Original Narrator)

Uses an older much simpler method to split text between Narrator and Character. It may also not filter unwanted characters as well as the newer version, resulting in more odd sounds produced by the TTS engine.

This is being left here as a fallback version, should anyone ever need it.