Maybe a bug in config.json
#1
by
WindyLee104 - opened
The configuration file (config.json) of this expert model named alpha-ai/Medical-Guide-COT-llama3.2-1B is completely wrong. It claims in the file that it is a model of Llama 2 specification (32000/4096), but its actual weight file is made according to Llama 3.21B (128256/2048) specification.
Just tested, its working. Maybe something off with the env.