araleza

I've noticed something similar too. I found a good 4 bit GPTQ model and I trained a LoRA for it, which worked great. I wanted to see what the unquantized...

Although one thing worth noting is that LoRAs are not great at training to add knowledge (such as 'who Sona is'). They mostly do style and personality. Using the character...

I noticed today that if I apply a LoRA to a 4-bit GPTQ model loaded with the Transformers model loader, it has no effect. But if I load that same...

Okay, so I started learning a bit about LoRA coding to try to fix this. I definitely haven't fixed it yet, but I have found a workaround (in case anyone...

Yep, that's what happens to me too. My issue for it is here: https://github.com/oobabooga/text-generation-webui/issues/5239 There's a workaround in that thread there to hand install yaml so it works for you....

I worked around it by forcing installation myself by doing: ``` ./cmd_linux.sh pip install -r requirements.txt ``` but that now leads to: ``` ERROR: Exception in ASGI application Traceback (most...

@Flowrome , if you've got the first bug (yaml not being found), you can probably work around that on Windows by typing: ``` cmd_windows.bat pip install -r requirements.txt ``` If...

Yes, @WhisperingDROID , but before you install it, make sure you activate the correct virtual environment by running cmd_linux.sh / cmd_windows.bat / cmd_macos.sh

I wanted to confirm, I had this yaml issue, and I was also running the installation through a symlink.

This sounds good, it would solve the 'change number of Mixtral experts' issue raised here: https://github.com/oobabooga/text-generation-webui/discussions/5367