Text Gestalt

#单卡训练（训练周期长，不建议）
python3 tools/train.py -c configs/sr/sr_tsrn_transformer_strock.yml

#多卡训练，通过--gpus参数指定卡号
python3 -m paddle.distributed.launch --gpus '0,1,2,3'  tools/train.py -c configs/sr/sr_tsrn_transformer_strock.yml

# GPU 评估， Global.pretrained_model 为待测权重
python3 -m paddle.distributed.launch --gpus '0' tools/eval.py -c configs/sr/sr_tsrn_transformer_strock.yml -o Global.pretrained_model={path/to/weights}/best_accuracy

# 预测使用的配置文件必须与训练一致
python3 tools/infer_sr.py -c configs/sr/sr_tsrn_transformer_strock.yml -o Global.pretrained_model={path/to/weights}/best_accuracy Global.infer_img=doc/imgs_words_en/word_52.png

python3 tools/export_model.py -c configs/sr/sr_tsrn_transformer_strock.yml -o Global.pretrained_model={path/to/weights}/best_accuracy Global.save_inference_dir=./inference/sr_out

python3 tools/infer/predict_sr.py --sr_model_dir=./inference/sr_out --image_dir=doc/imgs_words_en/word_52.png --sr_image_shape=3,32,128

@inproceedings{chen2022text,
  title={Text gestalt: Stroke-aware scene text image super-resolution},
  author={Chen, Jingye and Yu, Haiyang and Ma, Jianqi and Li, Bin and Xue, Xiangyang},
  booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
  volume={36},
  number={1},
  pages={285--293},
  year={2022}
}