how to export to Onnx model for faster inference

[eisneim]

this is a very useful project, would be great if it can be used in production with onnx support

[invictus717]

Thank you very much for your constructive suggestions! Could you please specify which part we need to provide onnx support first? Should it be the inference part of our multimodal encoder?

[eisneim]

Thank you! 👍 shared encoder for Text, Image, Audio, Video is the most useful one, those embedings can be used for LLM app that does audio, video retrieval; searching task would be: image to audio search, image to video clip search, text to audio search, etc.