kex0

Using `gpt4-x-alpaca-13b-native-4bit-128g` from https://huggingface.co/anon8231489123/gpt4-x-alpaca-13b-native-4bit-128g/tree/main Response to "hi": ``` . when-­ during on in on but... while............................................... trust.............................................................................................­............................h.............. ``` Or this when I replace these files `tokenizer_config.json` `tokenizer.model` `special_tokens_map.json` `generation_config.json` with...

Sadly the issue persists with `gpt4-x-alpaca-13b-native-4bit-128g` in the https://github.com/oobabooga/text-generation-webui/commit/a3085dba073fe8bdcfb5120729a84560f5d024c3 commit 

> Have you tried removing the non-cuda .pt file out of the directory and only having the cuda version? That solved the gibberish for me. Oh my.. I though I...

# Video [](https://www.youtube.com/watch?v=wy4AZ49wOF0)

I've been trying to keep this uptodate for a week without any word from maintainers. I'm only going to fix it once I know they want this kind of feature.

Doesn't look faked to me. Tons of artifacts and movements that are missing or are different from the original footage. Mainly faces and movement of clothes.

> Oh I see what you mean. I find it difficult to compare images side by side when I'm looking at the pixels. > > I was hoping for something...

Sadly I won't be able to fix this until I come back from vacation :(

> Thanks for the comments about switching to mediapipe. That does seem to make it work again. Really wish this extension was still maintained, as it works quite differently than...

Hey, yes, this project is dead because I think there are better ways of face-swapping like this https://github.com/Gourieff/sd-webui-reactor or if you still want to use stable diffusion to generate the...