Instance-Aware Dynamic Neural Network Quantization, CVPR 2022. Credits for Huawei Noah's Lab

Pytorch implementation of instance aware dynamic neural network quantization.

Key Features

How To Use

Usage

To train the model:

python3 eval.py --num-classes 200 --tar_bit 4 --train_dataset_path ./train

To evaluate the model:

python3 eval.py --val_dataset_path ./organized_val_data --num-classes 200 --tar_bit 4

Download

The dataset used is TinyImageNet

Emailware

Credits

Huawei Noah's Lab

Related

• Instance-Aware Dynamic Neural Network Quantization, CVPR 2022
• CADyQ

Support

Or

You may also like...

License

Apache 2.0

---