[Question]: Only the full version image can make use of GPU during parsing, not slim

[ChanningZhang]

Self Checks

• I have searched for existing issues search for existing issues, including closed ones.
• I confirm that I am using English to submit this report (Language Policy).
• Non-english title submitions will be closed directly ( 非英文标题的提交将会被直接关闭 ) (Language Policy).
• Please do not modify this template :) and fill in all the required fields.

Describe your problem

I tried v0.15.1, v0.16.0, v0.17.0 with same docker config, and found that only the full-version will take advantage of GPU during chunking, the slim-version will use CPU only, why? The gap between two versions should be the embedding models only (according to the Docs) right?

I set Qwen as embedding model, not the built-in ones.

[KevinHuSh]

The slim did not have pytorch, and that's why it's slim.

[ChanningZhang]

The slim did not have pytorch, and that's why it's slim.

Thanks for the clarification.