This is an un-official implementation of Paper An Improved One millisecond Mobile Backbone, whose performance is close to the paper.
Actually I achieve the MobileOne' s0 architecture (the smallest one), and validate on ImageNet-1000k dataset. And a val accuracy is here.
| Model | before merging blocks | after merging blocks | FLOPS |
|---|---|---|---|
| origin paper (s0) | none | 71.4 | 275M |
| my implementation (s0) | 70.470 | 70.518 | 274M |
Note that I only train a "s0" version, but you can easily modify the code to train other version, please see "mobileone.py" to refer to a configuration.
Different from original paper, We don't use:
- AutoAugment. In fact, S0 indeed abandons autoaugment.
- annealed weight decay. I set it constantly 4e-5.
- label smoothing regularization
- EMA update strategy
- progressive learning curriculum. I directly use 224px to train.
- Custom Weight decay Loss. I directly use WeightDecay in optimizer.
I release a pretrained model weight, click here to download. The test script validates the trained model, and also generates a converted deploy model.
python test.py {your imagenet-1000k dataset path} deploy mobileone_s0_hello_best.pth.tar
A converted deploy model is generated at "mobileone_deploy_model.pt" file.
I train the mobileone-s0 on 8 32G-V100 GPUS, costing about 4 days.
python train.py -a mobileone_s0 --dist-url 'tcp://127.0.0.1:23333' --dist-backend 'nccl' --multiprocessing-distributed --world-size 1 --rank 0 --workers 32 {your imagenet-1000k dataset path} --tag hello --wd 4e-5
model = make_mobileone_s0(deploy=True)
model.load(torch.load('mobileone_depoly_model.pt'))
model.cuda()
The entire code is based on RepVGG repository. Thanks for simply-using code.