Views
No views yet
| 模型 | 韩语数据集 精度 (%) |
|---|---|
| korean_PP-OCRv5_mobile_rec | 88.0 |
1# for CUDA11.8
2python -m pip install paddlepaddle-gpu==3.0.0 -i https://www.paddlepaddle.org.cn/packages/stable/cu118/
3
4# for CUDA12.6
5python -m pip install paddlepaddle-gpu==3.0.0 -i https://www.paddlepaddle.org.cn/packages/stable/cu126/
6
7# for CPU
8python -m pip install paddlepaddle==3.0.0 -i https://www.paddlepaddle.org.cn/packages/stable/cpu/python -m pip install paddleocr1paddleocr text_recognition \
2 --model_name korean_PP-OCRv5_mobile_rec \
3 -i https://cdn-uploads.huggingface.co/production/uploads/681c1ecd9539bdde5ae1733c/g-QlJbFcFy6VQ8hQJK3SD.jpeg1from paddleocr import TextRecognition
2model = TextRecognition(model_name="korean_PP-OCRv5_mobile_rec")
3output = model.predict(input="g-QlJbFcFy6VQ8hQJK3SD.jpeg", batch_size=1)
4for res in output:
5 res.print()
6 res.save_to_img(save_path="./output/")
7 res.save_to_json(save_path="./output/res.json"){'res': {'input_path': '/root/.paddlex/predict_input/g-QlJbFcFy6VQ8hQJK3SD.jpeg', 'page_index': None, 'rec_text': '리에 씨가 자기는돈이 많이있는다고 했어요', 'rec_score': 0.9142155051231384}}
1paddleocr ocr -i https://cdn-uploads.huggingface.co/production/uploads/681c1ecd9539bdde5ae1733c/rZK0SpYaWPbYsh2UBmRqV.png \
2 --text_recognition_model_name korean_PP-OCRv5_mobile_rec \
3 --use_doc_orientation_classify False \
4 --use_doc_unwarping False \
5 --use_textline_orientation True \
6 --save_path ./output \
7 --device gpu:0 1{'res': {'input_path': '/root/.paddlex/predict_input/rZK0SpYaWPbYsh2UBmRqV.png', 'page_index': None, 'model_settings': {'use_doc_preprocessor': True, 'use_textline_orientation': True}, 'doc_preprocessor_res': {'input_path': None, 'page_index': None, 'model_settings': {'use_doc_orientation_classify': False, 'use_doc_unwarping': False}, 'angle': -1}, 'dt_polys': array([[[ 15, 9],
2 ...,
3 [ 15, 27]],
4
5 ...,
6
7 [[ 7, 162],
8 ...,
9 [ 7, 187]]], dtype=int16), 'text_det_params': {'limit_side_len': 64, 'limit_type': 'min', 'thresh': 0.3, 'max_side_limit': 4000, 'box_thresh': 0.6, 'unclip_ratio': 1.5}, 'text_type': 'general', 'textline_orientation_angles': array([0, ..., 0]), 'text_rec_score_thresh': 0.0, 'rec_texts': ['보기', '저는오늘오후에 도서관에서 한국어 공부를 합니다영화를안 봅니다공부를 하기 전에 밥을', '먹습니다공부를한후에공원에서산책을 합니다민호 씨는백화점에서 쇼핑을 합니다오늘', '오후에 운동을 안 합니다.쇼핑을 하기 전에 은행에서 돈을 찾습니다쇼핑을 한 후에 식당에서', '밥을 먹습니다.'], 'rec_scores': array([0.99935752, ..., 0.96095514]), 'rec_polys': array([[[ 15, 9],
10 ...,
11 [ 15, 27]],
12
13 ...,
14
15 [[ 7, 162],
16 ...,
17 [ 7, 187]]], dtype=int16), 'rec_boxes': array([[ 15, ..., 27],
18 ...,
19 [ 7, ..., 187]], dtype=int16)}}save_path. The visualization output is shown below:
1from paddleocr import PaddleOCR
2
3ocr = PaddleOCR(
4 text_recognition_model_name="korean_PP-OCRv5_mobile_rec",
5 use_doc_orientation_classify=False, # Use use_doc_orientation_classify to enable/disable document orientation classification model
6 use_doc_unwarping=False, # Use use_doc_unwarping to enable/disable document unwarping module
7 use_textline_orientation=True, # Use use_textline_orientation to enable/disable textline orientation classification model
8 device="gpu:0", # Use device to specify GPU for model inference
9)
10result = ocr.predict("https://cdn-uploads.huggingface.co/production/uploads/681c1ecd9539bdde5ae1733c/rZK0SpYaWPbYsh2UBmRqV.png")
11for res in result:
12 res.print()
13 res.save_to_img("output")
14 res.save_to_json("output")PP-OCRv5_server_rec, and you can also use the local model file by argument text_recognition_model_dir. For details about usage command and descriptions of parameters, please refer to the Document.