PNASNet.TF

TensorFlow implementation of PNASNet-5. While completely compatible with the official implementation, this implementation focuses on simplicity and inference.

In particular, three files of 1200 lines in total (nasnet.py, nasnet_utils.py, pnasnet.py) are refactored into two files of 400 lines in total (cell.py, pnasnet.py). This code no longer supports NCHW data format, primarily because the released model was trained with NHWC. I tried to keep the rough structure and all functionalities of the official implementation when simplifying it.

If you use the code, please cite:

@inproceedings{liu2018progressive,
  author    = {Chenxi Liu and
               Barret Zoph and
               Maxim Neumann and
               Jonathon Shlens and
               Wei Hua and
               Li{-}Jia Li and
               Li Fei{-}Fei and
               Alan L. Yuille and
               Jonathan Huang and
               Kevin Murphy},
  title     = {Progressive Neural Architecture Search},
  booktitle = {European Conference on Computer Vision},
  year      = {2018}
}

Requirements

• TensorFlow 1.8.0
• torchvision 0.2.1 (for dataset loading)

Data and Model Preparation

• Download the ImageNet validation set and move images to labeled subfolders. To do the latter, you can use this script. Make sure the folder val is under data/.
• Download the PNASNet-5_Large_331 pretrained model:

cd data
wget https://storage.googleapis.com/download.tensorflow.org/models/pnasnet-5_large_2017_12_13.tar.gz
tar xvf pnasnet-5_large_2017_12_13.tar.gz

Usage

python main.py

The last printed line should read:

Test: [50000/50000]	[email protected] 0.829	[email protected] 0.962