Yuki Watanabe

Thank you for proposing this feature, @franco-bocci ! Adding index sounds valuable for the large scale use case of the model registry. However, one concern is the migration effort as...

Yes I think that works! Before actually start working on it, would you mind running a quick benchmark in your env with index/no-index, so we can get overall sense of...

@franco-bocci Yes that sounds like a plan! Please let us know when you get the benchmark (no time pressure!). Thank you so much for your willingness for contribution🙂

> The real issue is each client returns different objects. 100%, and I assume making change to Databricks client is too dangerous here? I think the solution [here](https://github.com/mlflow/mlflow/pull/11096#issuecomment-1970539657) is safer,...

Just found that the model exposes `dtype` property for getting dtype from parameter... ([source](https://github.com/huggingface/transformers/blob/main/src/transformers/modeling_utils.py#L883-L887)). Tho still used as fallback for missing model config in their source code. Should be simpler...

Made a change for dtype extracting logic, basically not to log it when the model's dtype is default one `torch.float32`. The issue is that `models.dtype` returns valid torch dtype even...

Btw you can rebase the PR to master to resolve the build_doc failure:)

Thanks for the review, @BenWilson2! > Verify that this builds in a windows environment Sure, do you know what is the easiest option to test this? remote desktop? > Create...

Using virtual Windows machine for testing was quite a bit of effort - permission setting, installing tools, etc. I ended up testing with my personal Windows laptop:p The basic `test_docker.py`...

@es94129 merged the PR, let's do few follow-ups; 1. We haven't implement the input preprocessing for `llm/v1/chat` task right? As mentioned in the previous note, you can simply add a...