snakers4 / MNASNet-pytorch-1

MNASNet implementation and pre-trained model in PyTorch

MNASNet in PyTorch

An implementation of MNASNet in PyTorch. MNASNet is an efficient convolutional neural network architecture for mobile devices, developed with architectural search. For more information check the paper: MnasNet: Platform-Aware Neural Architecture Search for Mobile

The model is is implemented by billhhh and the initial idea of reproducing MNASNet is by snakers4

Usage

Clone the repo:

git clone https://github.com/Randl/MNASNet-pytorch
pip install -r requirements.txt

Use the model defined in model.py to run ImageNet example:

python3 -m torch.distributed.launch --nproc_per_node=8 imagenet.py --dataroot "/path/to/imagenet/" --warmup 5 --sched cosine -lr 0.2 -b 128 -d 5e-5 --world-size 8 --seed 42

To continue training from checkpoint

python imagenet.py --dataroot "/path/to/imagenet/" --resume "/path/to/checkpoint/folder"

Results

Initially I've got 72+% top-1 accuracy, but the checkpointing didn't work properly. I believe the results are reproducable.

Classification Checkpoint	MACs (M)	Parameters (M)	Top-1 Accuracy	Top-5 Accuracy	Claimed top-1	Claimed top-5

You can test it with

python imagenet.py --dataroot "/path/to/imagenet/" --resume "results/shufflenet_v2_0.5/model_best.pth.tar" -e

Other implementations

• Mnasnet.MXNet -- A Gluon implementation of Mnasnet, 73.6% top-1 and 91.52% top-5
• MnasNet-pytorch-pretrained -- A PyTorch implementation of Mnasnet, 70.132% top-1 and 89.434% top-5