Open xyxxmb opened 3 weeks ago
output = torch.nn.functional.linear(A, F.dequantize_4bit(B, quant_state).to(A.dtype).t(), bias) RuntimeError: mat1 and mat2 shapes cannot be multiplied (332x4096 and 1x8388608)
same here, did u solve it?
bitsandbytes>0.43.1 will be ok,other version like 0.42 get worse
output = torch.nn.functional.linear(A, F.dequantize_4bit(B, quant_state).to(A.dtype).t(), bias) RuntimeError: mat1 and mat2 shapes cannot be multiplied (332x4096 and 1x8388608)