when i parser the pdf in "DataSet" , report the error;

[xuhongxin]

Begin at:
Mon, 03 Feb 2025 09:01:32 GMT
Duration:
235.86 s
Progress:
09:01:33 Task has been received.
09:01:34 Page(113): OCR started
09:01:51 Page(113): [ERROR]Internal server error while chunking: [ONNXRuntimeError] : 2 : INVALID_ARGUMENT : Did not find an arena based allocator registered for device-id combination in the memory arena shrink list: gpu:0
09:01:51 [ERROR][Exception]: [ONNXRuntimeError] : 2 : INVALID_ARGUMENT : Did not find an arena based allocator registered for device-id combination in the memory arena shrink list: gpu:0

[KevinHuSh]

pip unintall onnxruntime-gpu

[xuhongxin]

i run command with pip3 and shutdown the docker container and restart it ; but the it seems not work ; there still report the same error;Did I do something wrong?

[KevinHuSh]

What about startting ragflow server without GPU utilization?

[xuhongxin]

i use the commond

$ docker compose -f docker/docker-compose.yml up -d

how can i startting whitout GPU utilization?

[KevinHuSh]

Upgrade to the nightly version of docker image.