why TextRecognition crashed

[Mr-Glin]

model_dir is exists. software is running in podman container.but TextRecognition( xxx) fail, catch is not work.

[liuhongen1234567]

Hello, could you please specify what the exact error is? Can you isolate the TextRecognition class and run it separately to eliminate interference from other factors?

[Mr-Glin]

i can't catch error info .because it not into catch of try-catch .

[liuhongen1234567]

What I mean is to run the most basic text recognition code, for example:

from paddleocr import TextRecognition
model = TextRecognition(model_name="PP-OCRv4_mobile_rec", model_dir=XXX,cpu_threads=2)
output = model.predict(input="general_ocr_rec_001.png", batch_size=1)
for res in output:
    res.print()
    res.save_to_img(save_path="./output/")
    res.save_to_json(save_path="./output/res.json")

test whether it's an issue with PaddleOCR. Otherwise, we have no way to resolve it on our end.

[Mr-Glin]

cannot run to for res in output . crashed before TextRecognition(xxx)
I seem to have found a problem. The .paddlex I specified is empty. Is this the key to this problem?

[Mr-Glin]

Can it run normally without a network in the operating environment? If not, can it run without a network? For example, download the relevant files in advance.