dask-pytorch-ddp
dask-pytorch-ddp copied to clipboard
saturncloud
dask-pytorch-ddp is a Python package that makes it easy to train PyTorch models on dask clusters using distributed data parallel.
It would be great to expand data.S3ImageFolder to also support the following (in order of decreasing priority): - torchvision.datasets - torchtext.datasets - tensorflow datasets - GCP This would allow the...
It would be good to add the [`license_files` fields]( https://setuptools.readthedocs.io/en/latest/setuptools.html#metadata ) to [`setup`]( https://github.com/saturncloud/dask-pytorch/blob/d9d1645585f23c143521d2307703e125f9fd628f/setup.py#L14 ) to ensure downstream users both know the license of the software and receive the license...
Hi, When I was run this code (https://saturncloud.io/docs/examples/python/pytorch/qs-03-pytorch-gpu-dask-single-model/), I get this error: ```cmd daskcluster-worker-1 | 2022-11-13 17:01:17,386 - distributed.worker - WARNING - Compute Failed daskcluster-worker-1 | Key: dispatch_with_ddp-cbbbf432f092a3807b25cc40c48f7660 daskcluster-worker-1 |...
`dispatch.run` uses worker-restrictions to pin tasks to the workers they should get executed on. Should a worker get removed (or possibly restarted), this will cause the task to transition to...
Metadata
Owner
Metadata
dask-pytorch-ddp is a Python package that makes it easy to train PyTorch models on dask clusters using distributed data parallel.