DNA

This repository provides the code of our paper: Blockwisely Supervised Neural Architecture Search with Knowledge Distillation.

Illustration of DNA. Each cell of the supernet is trained independently to mimic the behavior of the corresponding teacher block.

Comparison of model ranking for DNA vs. DARTS, SPOS and MnasNet under two different hyper-parameters.

Our Trained Models

• Our searched models have been trained from scratch and can be found in: https://drive.google.com/drive/folders/1Oqc2gq8YysrJq2i6RmPMLKqheGfB9fWH.

• Here is a summary of our searched models:

  | Model | FLOPs | Params | [email protected] | [email protected] | |:---------:|:---------:|:---------:|:---------:|:---------:| | DNA-a | 348M | 4.2M | 77.1% | 93.3% | | DNA-b | 394M | 4.9M | 77.5% | 93.3% | | DNA-c | 466M | 5.3M | 77.8% | 93.7% | | DNA-d | 611M | 6.4M | 78.4% | 94.0% |

Usage

1. Requirements

• Install PyTorch (pytorch.org)
• Install third-party requirements

  • pip install timm==0.1.14

    We use this pytorch-image-models codebase to validate our models.
• Download the ImageNet dataset and move validation images to labeled subfolders
  • To do this, you can use the following script: https://raw.githubusercontent.com/soumith/imagenetloader.torch/master/valprep.shvalprep.sh
  • Only the validation set is needed in the evaluation process.

2. Searching

The code for supernet training, evaluation and searching is under

searching

cd searching

i) Train & evaluate the block-wise supernet with knowledge distillation

• Modify datadir in

  initialize/data.yaml

  to your ImageNet path.
• Modify nprocpernode in

  dist_train.sh

  to suit your GPU number. The default batch size is 64 for 8 GPUs, you can change batch size and learning rate in

  initialize/train_pipeline.yaml

• By default, the supernet will be trained sequentially from stage 1 to stage 6 and evaluate after each stage. This will take about 2 days on 8 GPUs with EfficientNet B7 being the teacher. Resuming from checkpoints is supported. You can also change

  start_stage

  in

  initialize/train_pipeline.yaml

  to force start from a intermediate stage without loading checkpoint.

• sh dist_train.sh

  ii) Search for the best architecture under constraint.

  Our traversal search can handle a search space with 6 ops in each layer, 6 layers in each stage, 6 stages in total. A search process like this should finish in half an hour with a single cpu. To perform search over a larger search space, you can manually divide the search space or use other search algorithms such as Evolution Algorithms to process our evaluated architecture potential files.

• Copy the path of architecture potential files generated in step i) to

  potential_yaml

  in

  process_potential.py

  . Modify the constraint in

  process_potential.py

  .

• python process_potential.py

  iii) Searching with multiple cells in each block.

  Please refer to the clarification from @MohanadOdema in this issue.

3. Retraining

The retraining code is simplified from the repo: pytorch-image-models and is under

retraining

• cd retraining

• Retrain our models or your searched models

  • Modify the

    run_example.sh

    : change data path and hyper-params according to your requirements
  • Add your searched model architecture to

    model.py

    . You can also use our searched and predefined DNA models.

  • sh run_example.sh

• You can evaluate our models with the following command:\

  python validate.py PATH/TO/ImageNet/validation --model DNA_a --checkpoint PATH/TO/model.pth.tar

  • PATH/TO/ImageNet/validation

    should be replaced by your validation data path.

  • --model

    :

    DNA_a

    can be replaced by

    DNA_b

    ,

    DNA_c

    ,

    DNA_d

    for our different models.

  • --checkpoint

    : Suggest the path of your downloaded checkpoint here.