KV Cached not working in MLX

#6
by statisticalplumber - opened

Model takes all messages to process everytime, ignoring cache.
Other models are working as expected.

Sign up or log in to comment