(self, tensor, scales, zeros)
| 183 | return q_tensor |
| 184 | |
| 185 | def dequant(self, tensor, scales, zeros): |
| 186 | tensor = (tensor - zeros) * scales |
| 187 | return tensor |
| 188 | |
| 189 | def dequant(self, tensor, scales, out_dtype=torch.bfloat16): |
| 190 | tensor_f = tensor.to(torch.float32) |