Transformers
GGUF
English
Inference Endpoints

Model seems broken

#2
by Utochi - opened

I'm excited to see this model functional in the future but what seems to break it is when it starts loading context. for example a character card of 900 tokens gets about 50% loaded before stalling out and breaking.
Using Faraday as my GUI and havent tested on other platforms.

This is a 16x MoE, are you sure your engine supports this? If its just stalling, its almost certainly an engine bug, not a model problem.

mradermacher changed discussion status to closed

ive switched over to using a different GUI and so it will load but just spits out gibberish sadly. I am experimenting with other GUIs and presets to see if i can get it to work, i really want to see what this has to offer

Sign up or log in to comment