lcsouzamenezes/ai-voice-cloning
1
1
Fork 0
forked from mrq/ai-voice-cloning
Code Pull Requests Activity

I forgot about the changelog and never kept up with it, so I'll just not use a changelog

Browse Source
This commit is contained in:
mrq 2023-03-08 05:14:50 +00:00
parent ff07f707cb
commit d7e75a51cf
2 changed files with 7 additions and 17 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

19
README.md
View File

@ -12,23 +12,8 @@ This is not endorsed by [neonbjb](https://github.com/neonbjb/). I do not expect
## Documentation ## Documentation
Please consult [the wiki](https://git.ecker.tech/mrq/ai-voice-cloning/wiki) for the documentation, including how to install, prepare voices for, and use the software. Please consult [the wiki](https://git.ecker.tech/mrq/ai-voice-cloning/wiki) for the documentation.
## Bug Reporting ## Bug Reporting
If you run into any problems, please refer to the [issues you may encounter](https://git.ecker.tech/mrq/ai-voice-cloning/wiki/Issues) wiki page first. Please don't hesitate to submit an issue. If you run into any problems, please refer to the [issues you may encounter](https://git.ecker.tech/mrq/ai-voice-cloning/wiki/Issues) wiki page first.
## Changelogs
Below will be a rather-loose changelogss, as I don't think I have a way to chronicle them outside of commit messages:
### `2023.02.22`
* greatly reduced VRAM consumption through the use of [TimDettmers/bitsandbytes](https://github.com/TimDettmers/bitsandbytes)
* cleaned up section of code that handled parsing output from training script
* added button to reconnect to the training script's output (sometimes skips a line to update, but it's better than nothing)
* actually update submodules from the update script (somehow forgot to pass `--remote`)
### `Before 2023.02.22`
Refer to commit logs.

5
src/train.py
View File

@ -21,9 +21,14 @@ if __name__ == "__main__":
args = parser.parse_args() args = parser.parse_args()
args.opt = " ".join(args.opt) # absolutely disgusting args.opt = " ".join(args.opt) # absolutely disgusting
with open(args.opt, 'r') as file: with open(args.opt, 'r') as file:
opt_config = yaml.safe_load(file) opt_config = yaml.safe_load(file)
if "WORLD_SIZE" in os.environ:
if int(os.environ["WORLD_SIZE"]) > 1 and opt_config["steps"]["gpt_train"]["optimizer"] == "adamw":
opt_config["steps"]["gpt_train"]["optimizer"] = "adamw_zero"
if "ext" in opt_config and "bitsandbytes" in opt_config["ext"] and not opt_config["ext"]["bitsandbytes"]: if "ext" in opt_config and "bitsandbytes" in opt_config["ext"] and not opt_config["ext"]["bitsandbytes"]:
os.environ['BITSANDBYTES_OVERRIDE_LINEAR'] = '0' os.environ['BITSANDBYTES_OVERRIDE_LINEAR'] = '0'
os.environ['BITSANDBYTES_OVERRIDE_EMBEDDING'] = '0' os.environ['BITSANDBYTES_OVERRIDE_EMBEDDING'] = '0'