Danny Friar
I have the following Luigi DAG in `luigi_external.py`. ``` import time import luigi class FileTask(luigi.ExternalTask): filename = luigi.Parameter() retry_count = 10 def output(self): return luigi.LocalTarget(self.filename) class DependsFileTask(luigi.Task): filename = luigi.Parameter()...
To enable features such as elastic training. Looks like a very useful framework for distributed deep learning and currently has integrations with Slurm, K8s, Ray etc. It would be interesting...
Setting up Dask and Ray on Armada is more work than on K8s due to not being able to use published helm charts to deploy clusters. Being able to do...
Node discovery is currently implemented by either having nodes coordinate via a distributed filesystem or otherwise waiting until the jobs are running and using the K8s API to check which...
The memory usage is much higher as a result of evaluating the kernel on this line. Is there any reason for not using the parent class method to do this?
The new uv tool (https://astral.sh/blog/uv) is *much* faster than pip in a lot of cases. Would it be possible to add support to the hatch environment building to allow using...
### Describe the bug ```bash ~> wandb import mlflow --mlflow-tracking-uri ... --target-project test-mlflow --target-entity test-mlflow ``` ```shell Traceback (most recent call last): File "/home/admin/.virtualenvs/research/bin/wandb", line 8, in sys.exit(cli()) File "/home/admin/.virtualenvs/research/lib/python3.10/site-packages/click/core.py",...
# Bug Report ### Describe the bug Exporting a torch model with a `clip` operation results in different behaviour with NaN values when inferencing with ONNX CPU vs GPU. ###...
┆Issue is synchronized with this [Jira Task](https://gr-oss.atlassian.net/browse/FML-65) by [Unito](https://www.unito.io)
Write and read to arrow/feather format ┆Issue is synchronized with this [Jira Task](https://gr-oss.atlassian.net/browse/FML-66) by [Unito](https://www.unito.io)