GPTQ pseudo-quantization saved weights (pt format) How load Re-evaluation #50
Description
GPTQ pseudo-quantization saved weights (pt format) How load re-evaluation, I set the --load parameter after the execution, found an error:
gptq-main\quant.py", line 636, in forward
raise ValueError('Only supports a single token currently.')