Quick question about your quantized Q8 model vs another

#1
by baesik - opened

Hello, quick question regarding the Q8 version. What would be the difference between your ~11GB quantized model and, say, this Q8 version? I assume it's because yours contains more modules and components for more workflows other than text encoders for image generation? I wasn't sure. Thanks!

Hello, quick question regarding the Q8 version. What would be the difference between your ~11GB quantized model and, say, this Q8 version? I assume it's because yours contains more modules and components for more workflows other than text encoders for image generation? I wasn't sure. Thanks!

Sounds about right.
You linked a "TE-only"-Version while mine is just the basic modell pulled from google and quantized using llama.cpp.

Hope this helps :)

Sign up or log in to comment