unexpected GPU and Debug warnings

[TLiley2]

🔎 Search before asking

• I have searched the PaddleOCR Docs and found no similar bug report.
• I have searched the PaddleOCR Issues and found no similar bug report.
• I have searched the PaddleOCR Discussions and found no similar bug report.

🐛 Bug (问题描述)

I keep getting unexpected GPU and Debug warnings whenever printing Detection + Recognition results

[2024/12/10 02:38:18] ppocr DEBUG: Namespace(help='==SUPPRESS==', use_gpu=True, use_xpu=False, use_npu=False, use_mlu=False, ir_optim=True, use_tensorrt=False, min_subgraph_size=15, precision='fp32', gpu_mem=500, gpu_id=0, image_dir=None, page_num=0, det_algorithm='DB', det_model_dir='C:\\Users\\tnu20/.paddleocr/whl\\det\\ml\\Multilingual_PP-OCRv3_det_infer', det_limit_side_len=960, det_limit_type='max', det_box_type='quad', det_db_thresh=0.3, det_db_box_thresh=0.6, det_db_unclip_ratio=1.5, max_batch_size=10, use_dilation=False, det_db_score_mode='fast', det_east_score_thresh=0.8, det_east_cover_thresh=0.1, det_east_nms_thresh=0.2, det_sast_score_thresh=0.5, det_sast_nms_thresh=0.2, det_pse_thresh=0, det_pse_box_thresh=0.85, det_pse_min_area=16, det_pse_scale=1, scales=[8, 16, 32], alpha=1.0, beta=1.0, fourier_degree=5, rec_algorithm='SVTR_LCNet', rec_model_dir='C:\\Users\\tnu20/.paddleocr/whl\\rec\\japan\\japan_PP-OCRv4_rec_infer', rec_image_inverse=True, rec_image_shape='3, 48, 320', rec_batch_num=6, max_text_length=25, rec_char_dict_path='C:\\Users\\tnu20\\anaconda3\\Lib\\site-packages\\paddleocr\\ppocr\\utils\\dict\\japan_dict.txt', use_space_char=True, vis_font_path='./doc/fonts/simfang.ttf', drop_score=0.5, e2e_algorithm='PGNet', e2e_model_dir=None, e2e_limit_side_len=768, e2e_limit_type='max', e2e_pgnet_score_thresh=0.5, 
e2e_char_dict_path='./ppocr/utils/ic15_dict.txt', e2e_pgnet_valid_set='totaltext', e2e_pgnet_mode='fast', use_angle_cls=False, cls_model_dir='C:\\Users\\tnu20/.paddleocr/whl\\cls\\ch_ppocr_mobile_v2.0_cls_infer', cls_image_shape='3, 48, 192', label_list=['0', '180'], cls_batch_num=6, cls_thresh=0.9, enable_mkldnn=False, cpu_threads=10, use_pdserving=False, warmup=False, sr_model_dir=None, sr_image_shape='3, 32, 128', sr_batch_num=1, draw_img_save_dir='./inference_results', save_crop_res=False, crop_res_save_dir='./output', use_mp=False, total_process_num=1, process_id=0, benchmark=False, save_log_path='./log_output/', show_log=True, use_onnx=False, return_word_box=False, output='./output', table_max_len=488, table_algorithm='TableAttn', table_model_dir=None, merge_no_span_structure=True, table_char_dict_path=None, formula_algorithm='LaTeXOCR', formula_model_dir=None, formula_char_dict_path=None, formula_batch_num=1, layout_model_dir=None, layout_dict_path=None, layout_score_threshold=0.5, layout_nms_threshold=0.5, kie_algorithm='LayoutXLM', ser_model_dir=None, re_model_dir=None, use_visual_backbone=True, ser_dict_path='../train_data/XFUND/class_list_xfun.txt', ocr_order_method=None, mode='structure', image_orientation=False, layout=True, table=True, formula=False, ocr=True, recovery=False, recovery_to_markdown=False, use_pdf2docx_api=False, invert=False, binarize=False, alphacolor=(255, 255, 255), lang='japan', det=True, rec=True, type='ocr', savefile=False, ocr_version='PP-OCRv4', structure_version='PP-StructureV2')
[2024/12/10 02:38:18] ppocr WARNING: The first GPU is used for inference by default, GPU ID: 0
[2024/12/10 02:38:21] ppocr WARNING: The first GPU is used for inference by default, GPU ID: 0
[2024/12/10 02:38:25] ppocr DEBUG: dt_boxes num : 4, elapsed : 0.45089125633239746
[2024/12/10 02:38:25] ppocr DEBUG: rec_res num  : 4, elapsed : 0.23304462432861328

🏃‍♂️ Environment (运行环境)

OS: Windows 10 Home 22H2
Environment: anaconda
Python: 3.11.2
PaddleOCR: 3.0.0b1
Install: pip
RAM: 8GB DDR4-3200 SO-DIMM
CPU: AMD Ryzen™ 9 4900HS Mobile Processor
CUDA: V12.6.85

🌰 Minimal Reproducible Example (最小可复现问题的Demo)

from paddleocr import PaddleOCR,draw_ocr
ocr = PaddleOCR(lang='japan') # need to run only once to download and load model into memory
img_path = 'PaddleOCR/doc/imgs_en/img_12.jpg'
result = ocr.ocr(img_path, cls=False)
for idx in range(len(result)):
    res = result[idx]
    for line in res:
        print(line)

[SWHL]

I think this is normal. I run the same code in the AI Studio, and the log is as following:

main.py

from paddleocr import PaddleOCR,draw_ocr
ocr = PaddleOCR(lang='japan') # need to run only once to download and load model into memory
img_path = 'img_12.jpg'
result = ocr.ocr(img_path, cls=False)
for idx in range(len(result)):
    res = result[idx]
    for line in res:
        print(line)

aistudio@jupyter-57084-6258410:~$ python main.py
[2024/12/10 18:30:15] ppocr DEBUG: Namespace(alpha=1.0, benchmark=False, beta=1.0, cls_batch_num=6, cls_image_shape='3, 48, 192', cls_model_dir='/home/aistudio/.paddleocr/whl/cls/ch_ppocr_mobile_v2.0_cls_infer', cls_thresh=0.9, cpu_threads=10, crop_res_save_dir='./output', det=True, det_algorithm='DB', det_box_type='quad', det_db_box_thresh=0.6, det_db_score_mode='fast', det_db_thresh=0.3, det_db_unclip_ratio=1.5, det_east_cover_thresh=0.1, det_east_nms_thresh=0.2, det_east_score_thresh=0.8, det_limit_side_len=960, det_limit_type='max', det_model_dir='/home/aistudio/.paddleocr/whl/det/ml/Multilingual_PP-OCRv3_det_infer', det_pse_box_thresh=0.85, det_pse_min_area=16, det_pse_scale=1, det_pse_thresh=0, det_sast_nms_thresh=0.2, det_sast_score_thresh=0.5, draw_img_save_dir='./inference_results', drop_score=0.5, e2e_algorithm='PGNet', e2e_char_dict_path='./ppocr/utils/ic15_dict.txt', e2e_limit_side_len=768, e2e_limit_type='max', e2e_model_dir=None, e2e_pgnet_mode='fast', e2e_pgnet_score_thresh=0.5, e2e_pgnet_valid_set='totaltext', enable_mkldnn=False, fourier_degree=5, gpu_id=0, gpu_mem=500, help='==SUPPRESS==', image_dir=None, image_orientation=False, ir_optim=True, kie_algorithm='LayoutXLM', label_list=['0', '180'], lang='japan', layout=True, layout_dict_path=None, layout_model_dir=None, layout_nms_threshold=0.5, layout_score_threshold=0.5, max_batch_size=10, max_text_length=25, merge_no_span_structure=True, min_subgraph_size=15, mode='structure', ocr=True, ocr_order_method=None, ocr_version='PP-OCRv4', output='./output', page_num=0, precision='fp32', process_id=0, re_model_dir=None, rec=True, rec_algorithm='SVTR_LCNet', rec_batch_num=6, rec_char_dict_path='/home/aistudio/.data/webide/pip/lib/python3.7/site-packages/paddleocr/ppocr/utils/dict/japan_dict.txt', rec_image_inverse=True, rec_image_shape='3, 48, 320', rec_model_dir='/home/aistudio/.paddleocr/whl/rec/japan/japan_PP-OCRv4_rec_infer', recovery=False, return_word_box=False, save_crop_res=False, save_log_path='./log_output/', scales=[8, 16, 32], ser_dict_path='../train_data/XFUND/class_list_xfun.txt', ser_model_dir=None, show_log=True, sr_batch_num=1, sr_image_shape='3, 32, 128', sr_model_dir=None, structure_version='PP-StructureV2', table=True, table_algorithm='TableAttn', table_char_dict_path=None, table_max_len=488, table_model_dir=None, total_process_num=1, type='ocr', use_angle_cls=False, use_dilation=False, use_gpu=True, use_mp=False, use_npu=False, use_onnx=False, use_pdf2docx_api=False, use_pdserving=False, use_space_char=True, use_tensorrt=False, use_visual_backbone=True, use_xpu=False, vis_font_path='./doc/fonts/simfang.ttf', warmup=False)
[2024/12/10 18:30:24] ppocr DEBUG: dt_boxes num : 75, elapse : 2.5091543197631836
[2024/12/10 18:30:25] ppocr DEBUG: rec_res num  : 75, elapse : 1.4642219543457031
[[[448.0, 179.0], [1162.0, 179.0], [1162.0, 219.0], [448.0, 219.0]], ('ACKNOWＬＥDGＥMENTS', 0.8688492774963379)]
[[[399.0, 346.0], [460.0, 346.0], [460.0, 386.0], [399.0, 386.0]], ('We', 0.9998581409454346)]
[[[469.0, 349.0], [582.0, 344.0], [583.0, 382.0], [470.0, 387.0]], ('Would', 0.9906611442565918)]
[[[594.0, 342.0], [661.0, 347.0], [658.0, 387.0], [591.0, 382.0]], ('like', 0.8581578135490417)]
[[[663.0, 348.0], [707.0, 348.0], [707.0, 384.0], [663.0, 384.0]], ('t0', 0.7024626731872559)]
[[[718.0, 346.0], [822.0, 346.0], [822.0, 383.0], [718.0, 383.0]], ('thank', 0.9994527697563171)]
[[[831.0, 345.0], [875.0, 345.0], [875.0, 383.0], [831.0, 383.0]], ('all', 0.8864550590515137)]
[[[884.0, 345.0], [947.0, 345.0], [947.0, 384.0], [884.0, 384.0]], ('the', 0.9983473420143127)]
[[[958.0, 344.0], [1129.0, 350.0], [1128.0, 393.0], [957.0, 387.0]], ('designers', 0.9998689293861389)]
[[[1139.0, 348.0], [1209.0, 348.0], [1209.0, 388.0], [1139.0, 388.0]], ('and', 0.999459445476532)]
[[[636.0, 395.0], [721.0, 391.0], [723.0, 432.0], [639.0, 437.0]], ('who', 0.834391176700592)]
[[[735.0, 398.0], [819.0, 398.0], [819.0, 431.0], [735.0, 431.0]], ('have', 0.9993776082992554)]
[[[833.0, 392.0], [926.0, 397.0], [924.0, 435.0], [831.0, 430.0]], ('been', 0.9997546076774597)]
[[[402.0, 402.0], [626.0, 400.0], [627.0, 431.0], [403.0, 433.0]], ('Contributors', 0.9941883087158203)]
[[[938.0, 400.0], [1088.0, 400.0], [1088.0, 431.0], [938.0, 431.0]], ('Involved', 0.9485199451446533)]
[[[1101.0, 400.0], [1136.0, 400.0], [1136.0, 433.0], [1101.0, 433.0]], ('ín', 0.5144662857055664)]
[[[1145.0, 394.0], [1210.0, 399.0], [1207.0, 439.0], [1142.0, 433.0]], ('the', 0.9993351101875305)]
[[[400.0, 452.0], [602.0, 442.0], [604.0, 480.0], [402.0, 490.0]], ('Production', 0.9584988355636597)]
[[[616.0, 446.0], [658.0, 446.0], [658.0, 483.0], [616.0, 483.0]], ('0f', 0.9157159328460693)]
[[[667.0, 446.0], [737.0, 446.0], [737.0, 484.0], [667.0, 484.0]], ('this', 0.9961860179901123)]
[[[747.0, 448.0], [852.0, 448.0], [852.0, 484.0], [747.0, 484.0]], ('book;', 0.9900836944580078)]
[[[862.0, 447.0], [952.0, 442.0], [954.0, 480.0], [864.0, 485.0]], ('their', 0.997888445854187)]
[[[965.0, 445.0], [1208.0, 448.0], [1207.0, 485.0], [964.0, 481.0]], ('Contributions', 0.9959254860877991)]
[[[401.0, 500.0], [486.0, 500.0], [486.0, 533.0], [401.0, 533.0]], ('have', 0.9995792508125305)]
[[[499.0, 498.0], [591.0, 498.0], [591.0, 534.0], [499.0, 534.0]], ('been', 0.9991582632064819)]
[[[602.0, 496.0], [856.0, 496.0], [856.0, 538.0], [602.0, 538.0]], ('indispensable', 0.9981399178504944)]
[[[868.0, 498.0], [910.0, 498.0], [910.0, 534.0], [868.0, 534.0]], ('t0', 0.6804746389389038)]
[[[919.0, 496.0], [966.0, 496.0], [966.0, 534.0], [919.0, 534.0]], ('its', 0.9990286827087402)]
[[[978.0, 500.0], [1132.0, 500.0], [1132.0, 531.0], [978.0, 531.0]], ('Creation', 0.9995676875114441)]
[[[1150.0, 496.0], [1210.0, 501.0], [1207.0, 537.0], [1147.0, 532.0]], ('We', 0.9994856119155884)]
[[[395.0, 549.0], [510.0, 544.0], [511.0, 582.0], [397.0, 586.0]], ('Would', 0.9098127484321594)]
[[[520.0, 548.0], [595.0, 548.0], [595.0, 586.0], [520.0, 586.0]], ('also', 0.9971821308135986)]
[[[602.0, 546.0], [667.0, 546.0], [667.0, 586.0], [602.0, 586.0]], ('like', 0.8751622438430786)]
[[[677.0, 553.0], [716.0, 553.0], [716.0, 584.0], [677.0, 584.0]], ('to', 0.6922805905342102)]
[[[728.0, 553.0], [864.0, 553.0], [864.0, 590.0], [728.0, 590.0]], ('express', 0.9985255599021912)]
[[[875.0, 553.0], [934.0, 553.0], [934.0, 588.0], [875.0, 588.0]], ('Our', 0.9559428095817566)]
[[[944.0, 551.0], [1109.0, 540.0], [1112.0, 584.0], [947.0, 594.0]], ('gratitude', 0.9993082880973816)]
[[[1122.0, 550.0], [1158.0, 550.0], [1158.0, 586.0], [1122.0, 586.0]], ('to', 0.8712176084518433)]
[[[1169.0, 546.0], [1211.0, 546.0], [1211.0, 586.0], [1169.0, 586.0]], ('all', 0.7641189694404602)]
[[[397.0, 593.0], [465.0, 601.0], [461.0, 641.0], [392.0, 633.0]], ('the', 0.9991509318351746)]
[[[474.0, 602.0], [664.0, 594.0], [665.0, 632.0], [475.0, 640.0]], ('producers', 0.980047345161438)]
[[[676.0, 595.0], [737.0, 595.0], [737.0, 636.0], [676.0, 636.0]], ('for', 0.9184491038322449)]
[[[747.0, 596.0], [836.0, 596.0], [836.0, 634.0], [747.0, 634.0]], ('their', 0.9984073638916016)]
[[[847.0, 593.0], [1041.0, 593.0], [1041.0, 634.0], [847.0, 634.0]], ('invaluable', 0.9987874031066895)]
[[[1054.0, 598.0], [1211.0, 598.0], [1211.0, 639.0], [1054.0, 639.0]], ('Opinions', 0.9326464533805847)]
[[[393.0, 651.0], [465.0, 646.0], [468.0, 684.0], [396.0, 688.0]], ('and', 0.9979880452156067)]
[[[475.0, 650.0], [651.0, 653.0], [651.0, 685.0], [474.0, 681.0]], ('assistance', 0.9994794726371765)]
[[[663.0, 650.0], [870.0, 650.0], [870.0, 691.0], [663.0, 691.0]], ('throughout', 0.9996002316474915)]
[[[878.0, 650.0], [941.0, 650.0], [941.0, 684.0], [878.0, 684.0]], ('this', 0.9970624446868896)]
[[[952.0, 650.0], [1082.0, 650.0], [1082.0, 688.0], [952.0, 688.0]], ('Project.', 0.9890819787979126)]
[[[1094.0, 646.0], [1167.0, 646.0], [1167.0, 684.0], [1094.0, 684.0]], ('And', 0.999886691570282)]
[[[1172.0, 652.0], [1209.0, 652.0], [1209.0, 683.0], [1172.0, 683.0]], ('to', 0.8733319640159607)]
[[[396.0, 696.0], [461.0, 701.0], [458.0, 740.0], [393.0, 735.0]], ('the', 0.9997127056121826)]
[[[465.0, 701.0], [562.0, 705.0], [561.0, 743.0], [463.0, 739.0]], ('many', 0.9946987628936768)]
[[[568.0, 702.0], [681.0, 696.0], [683.0, 734.0], [570.0, 740.0]], ('others', 0.9751121401786804)]
[[[693.0, 700.0], [807.0, 700.0], [807.0, 736.0], [693.0, 736.0]], ('whose', 0.9556367993354797)]
[[[817.0, 705.0], [931.0, 705.0], [931.0, 734.0], [817.0, 734.0]], ('names', 0.9950763583183289)]
[[[934.0, 704.0], [994.0, 699.0], [997.0, 732.0], [936.0, 737.0]], ('are', 0.9981686472892761)]
[[[1003.0, 702.0], [1064.0, 702.0], [1064.0, 734.0], [1003.0, 734.0]], ('not', 0.9824116230010986)]
[[[1065.0, 698.0], [1212.0, 694.0], [1213.0, 731.0], [1066.0, 735.0]], ('Credited', 0.9878785610198975)]
[[[396.0, 748.0], [464.0, 748.0], [464.0, 788.0], [396.0, 788.0]], ('but', 0.9950106739997864)]
[[[469.0, 749.0], [561.0, 756.0], [558.0, 794.0], [466.0, 787.0]], ('have', 0.9998989105224609)]
[[[567.0, 752.0], [667.0, 752.0], [667.0, 789.0], [567.0, 789.0]], ('made', 0.9979819059371948)]
[[[677.0, 752.0], [815.0, 748.0], [816.0, 789.0], [678.0, 793.0]], ('specific', 0.9752350449562073)]
[[[822.0, 753.0], [919.0, 753.0], [919.0, 791.0], [822.0, 791.0]], ('input', 0.9872061014175415)]
[[[926.0, 753.0], [961.0, 753.0], [961.0, 784.0], [926.0, 784.0]], ('1n', 0.5861291885375977)]
[[[971.0, 750.0], [1036.0, 750.0], [1036.0, 784.0], [971.0, 784.0]], ('this', 0.9995046854019165)]
[[[1048.0, 750.0], [1146.0, 750.0], [1146.0, 788.0], [1048.0, 788.0]], ('book,', 0.9974652528762817)]
[[[1157.0, 753.0], [1213.0, 753.0], [1213.0, 788.0], [1157.0, 788.0]], ('we', 0.7877120971679688)]
[[[397.0, 799.0], [504.0, 804.0], [503.0, 843.0], [395.0, 839.0]], ('thank', 0.9987764358520508)]
[[[506.0, 803.0], [578.0, 803.0], [578.0, 845.0], [506.0, 845.0]], ('you', 0.992954432964325)]
[[[579.0, 803.0], [640.0, 803.0], [640.0, 843.0], [579.0, 843.0]], ('for', 0.994827926158905)]
[[[640.0, 807.0], [726.0, 807.0], [726.0, 845.0], [640.0, 845.0]], ('your', 0.9916614294052124)]
[[[735.0, 803.0], [929.0, 803.0], [929.0, 838.0], [735.0, 838.0]], ('Continuous', 0.9704574346542358)]
[[[940.0, 805.0], [1085.0, 805.0], [1085.0, 841.0], [940.0, 841.0]], ('Support.', 0.9692855477333069)]

[GreatV]

The warnings and logs you are encountering during PaddleOCR execution, such as DEBUG and WARNING messages related to GPU usage and other settings, are not necessarily indicative of errors, but they may point to potential issues in your configuration or environment. Here’s a detailed breakdown of the situation and potential resolutions:

---

1. Explanation of the Logs

Warnings:

• [2024/12/10 02:38:18] ppocr WARNING: The first GPU is used for inference by default, GPU ID: 0
  • This indicates that PaddleOCR is defaulting to GPU ID 0 for inference. If you intended to use a specific GPU, you’ll need to explicitly set it using the gpu_id parameter.

Debug Messages:

• Namespace Configuration Logs:

  • The extensive debug log shows all the configuration parameters used by PaddleOCR, such as the detection and recognition algorithms (DB, SVTR_LCNet), model directories, GPU settings, and other hyperparameters.
  • These logs are helpful for debugging. They confirm that the models and settings are loaded correctly.

• Performance Metrics:

  • Logs like dt_boxes num: 4, elapsed: 0.450891... and rec_res num: 4, elapsed: 0.233044... indicate the detection and recognition steps are working and provide timing metrics.

---

2. Common Causes of Warnings

1. GPU Usage Warnings:

   • If your GPU is used but low memory or TensorRT optimizations are not applied, warnings may appear. This could happen if:
     • CUDA or cuDNN installations are mismatched with PaddleOCR.
     • The gpu_mem parameter (gpu_mem=500) is too low for your model.

2. Debug Logs:

   • Debug logs are enabled by default (show_log=True). These logs are verbose but not harmful. They can be disabled to clean up the output.

3. Unexpected Delays or Freezes:

   • If detection is slower than expected or freezes during subsequent runs, it could be caused by:
     • GPU memory fragmentation.
     • Issues with TensorRT or other optimizations.

4. Model Path Issues:

   • Ensure that the paths to detection, recognition, and classification models (e.g., det_model_dir, rec_model_dir, cls_model_dir) are correct.

---

3. Steps to Resolve

Step 1: Verify GPU Compatibility

• Check whether PaddlePaddle is compiled with CUDA:

  import paddle
  print(paddle.is_compiled_with_cuda())

  • If False, you need to install the GPU version of PaddlePaddle.

• Ensure the CUDA and cuDNN versions match your PaddlePaddle version:

  nvcc --version
  nvidia-smi

  • For PaddleOCR 3.0.0b1, you should use CUDA 11.2 or higher with compatible cuDNN.

Step 2: Increase GPU Memory Allocation

• The current gpu_mem setting is 500 MB, which might be too low for models like PP-OCRv4. Increase it to 1024 MB or higher:

  ocr = PaddleOCR(use_gpu=True, gpu_mem=1024)

Step 3: Disable Debug Logs

• To suppress debug logs, set show_log=False:

  ocr = PaddleOCR(show_log=False)

Step 4: Use TensorRT for Optimization (Optional)

• If your environment supports TensorRT, enable it for better performance:

  ocr = PaddleOCR(use_tensorrt=True)

Step 5: Explicitly Specify GPU ID

• If you have multiple GPUs, ensure PaddleOCR is using the intended one by setting gpu_id:

  ocr = PaddleOCR(use_gpu=True, gpu_id=0)

Step 6: Check Model Paths

• Verify that the model directories point to valid paths and contain the appropriate inference files (model.pdmodel, model.pdiparams, etc.):

  det_model_dir = 'C:\\Users\\tnu20/.paddleocr/whl\\det\\ml\\Multilingual_PP-OCRv3_det_infer'
  rec_model_dir = 'C:\\Users\\tnu20/.paddleocr/whl\\rec\\japan\\japan_PP-OCRv4_rec_infer'
  cls_model_dir = 'C:\\Users\\tnu20/.paddleocr/whl\\cls\\ch_ppocr_mobile_v2.0_cls_infer'

---

4. Additional Troubleshooting

Related Discussions

• GPU freezes during second inference:
  • Similar issues with GPU performance were discussed. A potential fix is to set use_tensorrt=True or reduce the batch size (rec_batch_num=1).

Related Issues

• Illegal Instruction Core Dump:
  • If you encounter crashes, ensure your CPU supports AVX instructions if you're running on CPU, or recompile PaddleOCR for your architecture.

---

5. Cleaned Configuration Recommendation

Here’s a cleaned version of your configuration for improved stability:

ocr = PaddleOCR(
    use_gpu=True,
    gpu_id=0,
    gpu_mem=1024,  # Increased memory allocation
    use_tensorrt=False,  # Enable if TensorRT is set up
    rec_batch_num=4,  # Reduce batch size if memory issues occur
    show_log=False  # Disable debug logs for cleaner output
)

---

6. Conclusion

By following the steps above, you should be able to resolve the unexpected warnings and improve the stability and performance of your PaddleOCR environment. If the issue persists, consider sharing a minimal reproducible example and double-checking your CUDA/cuDNN setup.

---

Response generated by 🤖 feifei-bot | chatgpt-4o-latest