module load anaconda # Server Only
conda create -n mmdl python=3.8
conda activate mmdl
conda install pytorch torchvision torchaudio pytorch-cuda=11.7 -c pytorch -c nvidia
pip install -U openmim
mim install mmcv-full
git clone https://github.com/open-mmlab/mmclassification.git
cd mmclassification
pip install -v -e .
pip install tensorboardpython tools/train.py configs/mobilenet_v3/mobilenet_v3_small_b128_cifar10.pysbatch tools/server_train.sh configs/mobilenet_v3/mobilenet_v3_small_b128_cifar10.pypython tools/analysis_tools/analyze_logs.py plot_curve work_dirs/xxx directory/xxx.log.json --keys train_accuracy accuracy_top-1 --title "xxx" --legend train val --out xxx.jpg
# Example
python tools/analysis_tools/analyze_logs.py plot_curve work_dirs/mobilenet_v3_large_b128_cifar100/20221120_045648.log.json --keys train_accuracy accuracy_top-1 --title "Baseline MobileNetV3_Large on CIFAR100" --legend train val --out baseline_cifar100.jpg - mobilenetv3_large_b128_cifar10:
- Accuracy: 94.02
- Logs: work_dirs/mobilenet_v3_large_b128_cifar10/20221119_224117.log
- mobilenet_v3_large_b128_cifar100:
- Accuracy: 75.47
- Logs: work_dirs/mobilenet_v3_large_b128_cifar100/20221110_222220.log
- cutmix_mobilenet_v3_large_b128_cifar10:
- Accuracy: 95.18999
- Logs: work_dirs/cutmix_mobilenet_v3_large_b128_cifar10/20221120_091955_cutmix.log
- cutmix_mobilenet_v3_large_b128_cifar100:
- Accuracy: 79.5400
- Logs: work_dirs/cutmix_mobilenet_v3_large_b128_cifar100/20221116_165504_cutmix.log
- poly_warmup5_mobilenet_v3_large_b128_cifar10:
- Accuracy: 94.539
- Logs: work_dirs/poly_warmup5_mobilenet_v3_large_b128_cifar10/poly_warmup5.log
- poly_warmup5_mobilenet_v3_large_b128_cifar100:
- Accuracy: 75.50999
- Logs: work_dirs/poly_warmup5_mobilenet_v3_large_b128_cifar100/poly_warmup5_large.log
- bsconvs_mobilenet_v3_large_b128_cifar10:
- Accuracy: 94.12
- Logs: work_dirs/bsconvs_mobilenet_v3_large_b128_cifar10/20221120_141652.log
- bsconvs_mobilenet_v3_large_b128_cifar100:
- Accuracy: 76.41
- Logs: work_dirs/bsconvs_mobilenet_v3_large_b128_cifar100/20221120_120748.log
-
1.cifar10_data_with_bsconv
- Accuracy:94.82999
- Logs:work_dirs/1.cifar10_data_with_bsconv/20221120_160426_cifar10_bsconv.log
-
2.cifar10_data_with_poly
- Accuracy:95.61
- Logs:work_dirs/2.cifar10_data_with_poly/20221120_235117_cifar10_poly.log
-
3.cifar10_bsconv_with_poly
- Accuracy:94.61
- Logs:work_dirs/3.cifar10_bsconv_with_poly/20221120_235711.log
-
4.cifar10_all_together
- Accuracy:95.28
- Logs:work_dirs/4.cifar10_all_together/20221121_024031.log
-
5.cifar100_data_with_bsconv
- Accuracy:79.45
- Logs:work_dirs/5.cifar100_data_with_bsconv/20221121_000725.log
-
6.cifar100_data_with_poly
- Accuracy:80.24
- Logs:work_dirs/6.cifar100_data_with_poly/20221121_021603.log
-
7.cifar100_bsconv_with_poly
- Accuracy:77.81
- Logs:work_dirs/7.cifar100_bsconv_with_poly/20221121_000402.log
-
8.cifar100_all_together
- Accuracy:80.9
- Logs:work_dirs/8.cifar100_all_together/20221121_040730.log