Enrico Shippole

> Indeed. This is something @kevinzwang is working on stabilizing our solution for. > > You can actually try it out with the environment variable: `DAFT_ENABLE_ACTOR_POOL_PROJECTIONS =1` > > The...

Hi @jaychia, Setting `DAFT_ENABLE_ACTOR_POOL_PROJECTIONS=1` helped to resolve the issue of the model loading multiple times per completed partition. But now when I run the above udf with : ```python STEmbeddingUDF.with_init_args(...

> Oops yes, I think you've stumbled upon one of the TODOs in this beta feature: > > Let me work on something here Thank you!

Hi @jaychia, I am now able to run with `EmbeddingUDF.with_init_args(model_name).with_concurrency(1).override_options(num_gpus=1)` using the main branch. But now hit an error for `RuntimeError: No CUDA GPUs are available` when using `daft.context.set_runner_ray()`. I...

> Hi @conceptofmind, just merged in a PR to fix the ray actor resource allocation issue you encountered. It will be included in our next release! Hi @kevinzwang, I will...

Looks like it is all working now on the main branch. See 1.0/1.0 in the Ray Actor pool: ``` 1.0/1.0 GPU ``` I will look into testing on multiple GPUs...

Yes will close out after testing multi gpu single node and multi GPU multi node. Thank you!

> [@conceptofmind](https://github.com/conceptofmind) how did you download the file? > > If i run the same code on the youtube url, it works as expected > > import daft > >...

> [@conceptofmind](https://github.com/conceptofmind) which version are you using, 0.6? > > The `read_video_frames` method will search all available formats to find the one closest to your desired resolution, then it will...

> could you provide the exact scripts/commands you used to download the files? I'm unable to reproduce locally. > > I also noticed that the specific video does not have...