What quantize method you use
GPTQ Int8 for weights.
Is this padded for multi-GPU use?
edit
· Sign up or log in to comment