torch.nn.quantized#
State collector class for float operations. |
|
module to replace FloatFunctional module before FX graph mode quantization, since activation_post_process will be inserted in top level module directly |
|
Wrapper class for quantized operations. |