FromS2B

TODO

• Upload a training/test datasets and checkpoints.
• Upload data preprocessing details.

Environment Setup

1. Clone Repository

git clone --recursive https://github.com/kmax2001/EvSharp2Blur.git
cd FromS2B

2. Create Environment

conda create -n froms2b python=3.8 -y
conda activate froms2b
pip install torch==1.8.1+cu111 torchvision==0.9.1+cu111 torchaudio==0.8.1 -f https://download.pytorch.org/whl/torch_stable.html
pip install -r requirements.txt

3. Prepare Datasets / External Resources

Train / Test by Stage

Stage 1

Train

python tools/train.py --cfg experiments/crowdpose/w32/w_32_train_event_subteacher.yaml

python tools/train.py --cfg experiments/crowdpose/w32/w_32_train_blur2blur_subteacher.yaml

python tools/train_step1.py --cfg experiments/crowdpose/w32/w_32_train_step1_blur2blur_subteacher.yaml

Test / Validation

python tools/valid_step1.py --cfg experiments/crowdpose/w32/w_32_test_event_subteacher.yaml

python tools/valid_step1.py --cfg experiments/crowdpose/w32/w_32_test_blur2blur_subteacher.yaml

python tools/valid_step1.py --cfg experiments/crowdpose/w32/w_32_test_step1_blur2blur_subteacher.yaml

Stage 2

Train

python tools/train_step2.py --cfg experiments/crowdpose/w32/w_32_train_step2.yaml

Test / Validation

python tools/valid_step2.py --cfg experiments/crowdpose/w32/w_32_test_step2.yaml

Stage 3

Train

python tools/train_step3.py --cfg experiments/crowdpose/w32/w_32_train_step3.yaml

Test / Validation

python tools/valid_step3.py --cfg experiments/crowdpose/w32/w_32_test_step3.yaml

Stage 4

Train

python tools/train_step4.py --cfg experiments/crowdpose/w32/w_32_train_step4.yaml

Test / Validation

python tools/valid_step4.py --cfg experiments/crowdpose/w32/w_32_test_step4.yaml