[replicate] support simpler device_id #100217
Conversation
Allow passing in `device_id=[device]` regardless of CPU or GPU. We modify the kwarg as needed to pass to DDP. [ghstack-poisoned]
rohan-varma
requested review from
mrshenli,
zhaojuanmao,
H-Huang,
awgu,
kwen2501,
wanchaol,
fegin,
yhcharles,
kiukchung and
d4l3k
as code owners
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/100217
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit 2954569: This comment was automatically generated by Dr. CI and updates every 15 minutes. |
Allow passing in `device_id=[device]` regardless of CPU or GPU. We modify the kwarg as needed to pass to DDP. ghstack-source-id: e2716d05b97be882ad05c6635f233ebf1e395bdb Pull Request resolved: #100217
This PR needs a labelIf your changes are user facing and intended to be a part of release notes, please use a label starting with If not, please add the To add a label, you can comment to pytorchbot, for example For more information, see |
| @@ -213,6 +214,49 @@ def test_replicate_with_kwargs(self): | |||
| ) | |||
| self._compare_module(model, replicate_model) | |||
|
|
|||
| def test_replicate_device_id(self): | |||
There was a problem hiding this comment.
AssertionError: Torch not compiled with CUDA enabled
| def test_replicate_device_id(self): | |
| @skip_if_lt_x_gpu(2) | |
| def test_replicate_device_id(self): |
| @@ -1,4 +1,5 @@ | |||
| from typing import Any, Dict, Iterable, List, Optional, Set, Tuple | |||
Allow passing in `device_id=[device]` regardless of CPU or GPU. We modify the kwarg as needed to pass to DDP. [ghstack-poisoned]
Allow passing in `device_id=[device]` regardless of CPU or GPU. We modify the kwarg as needed to pass to DDP. ghstack-source-id: 626353f80a6271f526f562fcc51d3ee308570e2f Pull Request resolved: #100217
|
@pytorchbot merge -f "CI done" |
Merge startedYour change will be merged immediately since you used the force (-f) flag, bypassing any CI checks (ETA: 1-5 minutes). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
Stack from ghstack (oldest at bottom):
Allow passing in
device_id=[device]regardless of CPU or GPU. Wemodify the kwarg as needed to pass to DDP.