Did anyone tried to work with post training method DFQ with lower bit widths such as 2-8? I am getting really low accuracies.
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
|
Bitwidth in AIMET PyTorch Quantization
|
6 | 1278 | March 7, 2021 | |
|
QAT for ResNet101 PyTorch & TensorFlow
|
0 | 718 | January 28, 2022 | |
|
Efficient-Net-B0 AIMET quantization
|
2 | 1152 | July 28, 2021 | |
|
Support PReLU or LeakyReLU?
|
0 | 664 | November 30, 2020 | |
|
Saving and loading compressed model
|
3 | 941 | November 20, 2020 |