Any plan for LangChain integration?

[tigerinus]

Would love to use model accelerated by Intel/OpenVINO in LangChain.

[devangaggarwal]

Thank you for this feedback and suggestion. We will definitely take this into account in our planning.

[nacartwright]

Ditto here too. I'm seeing customers being concerned about the token cost for using cloud models.

[yury-gorbachev]

This is good feature, I agree.
Is there any possibility to use this via HF adapter that they have and use optimum-intel for that I wonder?

[yury-gorbachev]

Meanwhile, you might want to check this: https://github.com/OpenVINO-dev-contest/langchain.openvino

We are considering upstreaming this, any feedback would be useful.

cc: @OpenVINO-dev-contest

[yury-gorbachev]

Just to keep this in the thread, update from @OpenVINO-dev-contest

OpenVINO integration with LangChain is ready, including:

• LLM: https://python.langchain.com/docs/integrations/llms/openvino
• Embedding: https://python.langchain.com/docs/integrations/text_embedding/openvino
• Rerank：https://python.langchain.com/docs/integrations/document_transformers/openvino_rerank