George

Unfortunately, it is not possible to use a custom json tool in pydantic v2 (e.g. https://github.com/pydantic/pydantic/discussions/6388) We've looked into the other places where we could replace the builtin json module...

Thank you for addressing this

hi @michaelfeil sorry for the late response, we'll review it in the closest time thank you for the contribution :)

hey @shahad-mahmud This feature is not stable yet, I'd recommend to disable it Once it is supported in qdrant, qdrant-client won't throw exceptions, but access to these metrics in qdrant-client...

hi @RaihanulHaque it's better to split your data in chunks you can do it manually and write a loop with `upsert` on your own or use convenience methods like `upload_collection`...

my comments are mostly nit, other than that looks good

Hi, yes, you can run fastembed models on gpu Have you seen this [example](https://github.com/qdrant/fastembed?tab=readme-ov-file#%EF%B8%8F-fastembed-on-a-gpu) ?

Hey @Pringled Sounds good, do you think you could submit a PR with a couple of the most prosperous models?

We would rather prefer to keep it onnx only What inputs does it have? Maybe we could adjust the interface or introduce a separate class

The fix should've been available as of qdrant-client 1.13.0, closing as completed