Skip to content

jiaowoguanren0615/SBCFormer-Pytorch

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 
 
 

Repository files navigation

SBCFormer-Pytorch

Paper:SBCFormer: Lightweight Network Capable of Full-size ImageNet Classification at 1 FPS on Single Board Computers(https://arxiv.org/abs/2311.03747)

Precautions

Note: Before training the classification model, you need to enter the train_gpu.py file to modify the data_root and batchsize and epochs of your own data set.The code turns on automatic mixed precision by default. If your GPU(s) does not support automatic mixed precision (you can use torch.cuda.is_bf16_supported() to check), you need to modify it in the util.engine.py file.

Train multi-gpus

Run command: torchrun --nproc_per_node=8 train_gpu.py

Train with specify specific GPUs

Run command: CUDA_VISIBLE_DEVICES=1,3 torchrun --nproc_per_node=2 train_gpu.py

Train single-gpu

Run command: python train_gpu.py

About

This is a warehouse for SBCFormer-pytorch-model, can be used to train your dataset

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages