byaldi

byaldi copied to clipboard

When installing the package with ```pip install byaldi``` the integrations folder is not installed. The folder is indeed missing in the .tar.gz and .whl files on [PyPi](https://pypi.org/project/Byaldi/#files)

hubkrieb

Hello. I have the index saved to disk (4 json.gz files, and one embeddings_0.pt file). `from_index()` takes anywhere between 15-30 seconds, which seems like a long time, given that the...

kubni

Why === Currently if you pass in the `device` arg to `from_pretrained`, there is a logic bug which makes it get resolved to something else unnecessarily based on the device...

ms337

I downloaded the models manually, is byaldi does automatically is not working in my case (guess that I don't know where to put the hugging face token). So, where should...

metantonio

remove_from_index functionality

1

Hey Team, Can I expect removal of index functionality any time soon? I'm really in need of this.

adi7820

Currently, Byaldi requires local inference by `RAG = RAGMultiModalModel.from_pretrained("vidore/colpali-v1.2")`. Can you refactor the `RAGMultiModalModel` class to call a remote VLLM API to complete the work? E.g. `RAGMultiModalModel.from_api("https://localhost:3000/v1/completions")`.

thiner

Is this normal for 1 pdf file with 15 pages for `vidore/colqwen2-v0.1` model ```sh CUDA out of memory. Tried to allocate 59.39 GiB. GPU 0 has a total capacity of...

iamkhalidbashir

Model is not being offloaded from VRAM

3

I am trying to run the model in Jupyter notebook.  1. In the above iteration I haven't initialized the model.  2. Now I run the cell the model...

nishithshowri006

I have a microservice that exposes some endpoints to interact with the model (like vllm or similar). How can I use my external model with Byaldi?

sir3mat

For #30 I basically used Claude to refactor the classes, no breaking changes introduced to the existing api, RAGMultiModalModel will internally use IndexManager (can also be a wrapper over a...

jdchawla29