Vietnamses-science-text-recognition

Requirement

• torch 1.10.0
• torchvision 0.7.0
• mmcv-full 1.3.16
• mmdet 2.18.1
• mmocr 0.3.0
• opencv-python 4.5.4.60
• cudatoolkit 10.2.89

TEST MODEL

• python tools/test.py ${CONFIG_FILE} ${CHECKPOINT_FILE} [ARGS]
• For eval detection --eval : 'hmean-iou', OCR --eval : 'acc'

RUN PREDICT MODEL

• Inference.py (change file config path)

Train model

• python tools/train.py ${CONFIG_FILE} [ARGS]

Pretrained model

• Weight and some image visualization :Link Here