PaddleOCR 3.0重磅发布！OCR精度跃升13%，多场景文档解析全面升级

pip install paddleocr

from paddleocr import PaddleOCRocr = PaddleOCR(    use_doc_orientation_classify=False,    use_doc_unwarping=False,    use_textline_orientation=False)result = ocr.predict(    input="https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/general_ocr_002.png")for res in result:    res.print()    res.save_to_img("output")    res.save_to_json("output")

from pathlib import Pathfrom paddleocr import PPStructureV3pipeline = PPStructureV3(    use_doc_orientation_classify=False,    use_doc_unwarping=False)output = pipeline.predict(    input="https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/pp_structure_v3_demo.png")for res in output:    res.print()    res.save_to_json(save_path="output")    res.save_to_markdown(save_path="output")

from paddleocr import PPChatOCRv4Doc
chat_bot_config = {    "module_name": "chat_bot",    "model_name": "ernie-3.5-8k",    "base_url": "https://qianfan.baidubce.com/v2",    "api_type": "openai",    "api_key": "api_key",  # your api_key}
retriever_config = {    "module_name": "retriever",    "model_name": "embedding-v1",    "base_url": "https://qianfan.baidubce.com/v2",    "api_type": "qianfan",    "api_key": "api_key",  # your api_key}
pipeline = PPChatOCRv4Doc(    use_doc_orientation_classify=False,    use_doc_unwarping=False)
visual_predict_res = pipeline.visual_predict(    input="https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/vehicle_certificate-1.png",    use_common_ocr=True,    use_seal_recognition=True,    use_table_recognition=True,)
mllm_predict_info = Noneuse_mllm = False# If a multimodal large model is used, the local mllm service needs to be started. You can refer to the documentation: https://github.com/PaddlePaddle/PaddleX/blob/release/3.0/docs/pipeline_usage/tutorials/vlm_pipelines/doc_understanding.en.md performs deployment and updates the mllm_chat_bot_config configuration.if use_mllm:    mllm_chat_bot_config = {        "module_name": "chat_bot",        "model_name": "PP-DocBee",        "base_url": "http://127.0.0.1:8080/",  # your local mllm service url        "api_type": "openai",        "api_key": "api_key",  # your api_key    }
    mllm_predict_res = pipeline.mllm_pred(        input="https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/vehicle_certificate-1.png",        key_list=["驾驶室准乘人数"],        mllm_chat_bot_config=mllm_chat_bot_config,    )    mllm_predict_info = mllm_predict_res["mllm_res"]
visual_info_list = []for res in visual_predict_res:    visual_info_list.append(res["visual_info"])    layout_parsing_result = res["layout_parsing_result"]
vector_info = pipeline.build_vector(    visual_info_list, flag_save_bytes_vector=True, retriever_config=retriever_config)chat_result = pipeline.chat(    key_list=["驾驶室准乘人数"],    visual_info=visual_info_list,    vector_info=vector_info,    mllm_predict_info=mllm_predict_info,    chat_bot_config=chat_bot_config,    retriever_config=retriever_config,)print(chat_result)