[Bug]: Is it a bug of the new release v0.17.0 or just the embedding issue?

[qwweewwee]

Is there an existing issue for the same bug?

• I have checked the existing issues.

RAGFlow workspace code commit ID

v0.17.0

RAGFlow image version

v0.17.0

Other environment information

ubuntu 22.04
RAM 64G, GRAPHIC CARD 16G NVIDIA4080

Actual behavior

While parsing pdf files, it keep giving errors as below. I also checked the ragflow log, it raised some alerts as well. BTW, I am running the newest release v0.17.0 from source.

Expected behavior

No response

Steps to reproduce

running the newest version from source

Additional information

When I started the backend service, i.e. ragflow, some message got my attention . wondering it might have something to do with the issue.

[KevinHuSh]

This is out of memory issue. onnxruntion has that kind of issue.
You could use CPU for parsing and apply an Olllam service to do embedding with GPUs.

[qwweewwee]

This is out of memory issue. onnxruntion has that kind of issue. You could use CPU for parsing and apply an Olllam service to do embedding with GPUs.

how can i fix it? are there any configurations for that ?

[KevinHuSh]

Deploy an Ollama server to serve embedding model on GPU.

[qwweewwee]

Deploy an Ollama server to serve embedding model on GPU.

ollama server is already running on GPU，but how to use cpu (or both ) for parsing ?

[KevinHuSh]

docker compose up via default docker-compose.ymal.