The status of this repository?

[dequeueing]

Hi. GPTCache is cool, but there has been no commits for 5 months. Howeve readme mentions that "repository is still under heavy development".

Can anyone explain what is the situation now for the repo?

[SimFG]

Are you having any problems or have any other ideas? Let's discuss it together.

[dequeueing]

Are you having any problems or have any other ideas? Let's discuss it together.

OK. Is GPTCache integrated into Langchain? Readme mentions this but the link is dead.

🎉 GPTCache has been fully integrated with 🦜️🔗LangChain ! Here are detailed usage instructions.

[SimFG]

because the langchain doc link has changed, you can visit the langchain repo to find the doc. I think it's https://python.langchain.com/api_reference/community/cache/langchain_community.cache.GPTCache.html#gptcache

[dequeueing]

because the langchain doc link has changed, you can visit the langchain repo to find the doc. I think it's https://python.langchain.com/api_reference/community/cache/langchain_community.cache.GPTCache.html#gptcache

Thanks for the link.

I wonder is there other way to support locally deployed huggingface model (e.g. llama2) , rather than using API? Could you please suggest?

[SimFG]

Currently, GPTCache supports the APIs of OpenAI and Cohere. You can try to package your embedding service into an API that is compatible with them. Similar to how most LLM models are now compatible with OpenAI’s SDK.

[dequeueing]

Currently, GPTCache supports the APIs of OpenAI and Cohere. You can try to package your embedding service into an API that is compatible with them. Similar to how most LLM models are now compatible with OpenAI’s SDK.

Great. Thanks for the advice.

[dequeueing]

Currently, GPTCache supports the APIs of OpenAI and Cohere. You can try to package your embedding service into an API that is compatible with them. Similar to how most LLM models are now compatible with OpenAI’s SDK.

Sorry I may not have thoroughly undertstood your meaning. What is the "embedding service"? Is it the "embedding generator" in your paper?

[SimFG]

ref: https://github.com/zilliztech/GPTCache/blob/main/docs/usage.md#build-your-cache

[dequeueing]

ref: https://github.com/zilliztech/GPTCache/blob/main/docs/usage.md#build-your-cache

Thanks. I wonder how can I manage the cache such as delete some entry in the cache. I notice that there are no such APIs in api.py. From the doc the way to achieve seperation is via Session. But I want to know can I delete some entry in the cache?

Thanks!

[SimFG]

You can try to learn about eviction manager., link: https://github.com/zilliztech/GPTCache/blob/main/gptcache/manager/eviction_manager.py