recursionbane

Your links are broken.

I'm facing this issue right now, and I'm considering this strategy: - Use `drive md5sum -r` and dump to a file - Parse this file into a hashes-of-array data structure...

Thanks, @dawoodkhan82!

Yes, AzureOpenAI support, please.

Seeing the same error, is it the sheer number of documents that matters?

I reduced chunk size and overlap until it worked reliably.

It depends on the complexity of your corpus. I reduced to 400 chunk size and 40 overlap to get it to work.

Working out of the box: * ggml-gpt4all-j-v1.3-groovy.bin Working after changing `backend='llama'` on line 30 in `privateGPT.py`, [thanks](https://github.com/imartinez/privateGPT/issues/276#issuecomment-1554262627) to @PulpCattel: * ggml-vicuna-13b-1.1-q4_2.bin * ggml-gpt4all-l13b-snoozy.bin Fails due to "[bad magic](https://github.com/imartinez/privateGPT/issues/280)": * ggml-mpt-7b-instruct.bin

I would recommend supporting only local HF models and deferring support for local-server models to something like https://github.com/go-skynet/LocalAI, which conforms models to the OpenAI API.

I made some other changes like removing the limit of retries in this fork: https://github.com/recursionbane/google-photos-delete-tool