Awesome multilingual OCR toolkits based on PaddlePaddle (practical ultra lightweight OCR system, support 80+ languages recognition, provide data annotation and synthesis tools, support training and deployment among server, mobile, embedded and IoT devices)
🔎 Search before asking
🐛 Bug (问题描述)
🏃♂️ Environment (运行环境)
ubuntu24.2
🌰 Minimal Reproducible Example (最小可复现问题的Demo)
运行命令
python -m paddle.distributed.launch --gpus 0,1 tools/train.py -c /opt/PaddleOCR/configs/det/ch_PP-OCRv3/ch_PP-OCRv3_det_student.yml -o \ Global.pretrained_model=/opt/PaddleOCR/pretrain_models/ch_PP-OCRv3_det_distill_train/student.pdparams \ Global.save_model_dir=/opt/PaddleOCR/output/train/det \ Global.epoch_num=300 \ Global.eval_batch_step="[0, 772]" \ Global.save_epoch_step=10000 \ Optimizer.lr.name=Const \ Optimizer.lr.learning_rate=0.0005 \ Optimizer.lr.warmup_epoch=0 \ Train.dataset.data_dir=/opt/PaddleOCR/train_data \ Train.dataset.label_file_list="[/opt/PaddleOCR/train_data/2024092501/det/train.txt, /opt/PaddleOCR/train_data/2024092502/det/train.txt]" \ Train.dataset.ratio_list="[0.5, 0.5]" \ Train.loader.batch_size_per_card=16 \ Train.loader.num_workers=8 \ Eval.dataset.data_dir=/opt/PaddleOCR/train_data \ Eval.dataset.label_file_list="[/opt/PaddleOCR/train_data/2024092501/det/test.txt, /opt/PaddleOCR/train_data/2024092502/det/test.txt]" \ Eval.loader.batch_size_per_card=16 \ Eval.loader.num_workers=8