Skip to content

[OPT] Run test in lower precision on GPU #17353

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
patrickvonplaten merged 8 commits into from
May 19, 2022
Merged

[OPT] Run test in lower precision on GPU #17353

Changes from 1 commit
Commits
Show all changes
8 commits
Select commit Hold shift + click to select a range
db27c99
[OPT] Run test only in half precision
patrickvonplaten May 19, 2022
77b2f6e
up
patrickvonplaten May 19, 2022
a533e06
up
patrickvonplaten May 19, 2022
cccd907
up
patrickvonplaten May 19, 2022
6002a65
up
patrickvonplaten May 19, 2022
f72ae0d
finish
patrickvonplaten May 19, 2022
3db3ac2
fix on GPU
patrickvonplaten May 19, 2022
8f1cc48
Update tests/models/opt/test_modeling_opt.py
patrickvonplaten May 19, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 6 additions & 3 deletions tests/models/opt/test_modeling_opt.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -268,17 +268,20 @@ def _long_tensor(tok_lst):
@require_torch
class OPTModelIntegrationTests(unittest.TestCase):
@slow
@unittest.skipIf(torch_device == "cpu", "Cant do half precision")
def test_inference_no_head(self):
model = OPTModel.from_pretrained("facebook/opt-350m").to(torch_device)
model = OPTModel.from_pretrained("facebook/opt-350m", torch_dtype=torch.float16).to(torch_device)
input_ids = _long_tensor([[0, 31414, 232, 328, 740, 1140, 12695, 69, 46078, 1588, 2]])
with torch.no_grad():
output = model(input_ids=input_ids).last_hidden_state
expected_shape = torch.Size((1, 11, 512))
self.assertEqual(output.shape, expected_shape)
expected_slice = torch.tensor(
[[-0.2867, -1.9256, -0.3062], [-1.2711, -0.1337, -0.1897], [0.4109, 0.1187, -1.3142]], device=torch_device
[[-0.2878, -1.9219, -0.3018], [-1.2588, -0.1310, -0.1711], [0.4131, 0.1089, -1.2939]],
device=torch_device,
dtype=torch.float16,
)
self.assertTrue(torch.allclose(output[:, :3, :3], expected_slice, atol=1e-3))
assert_tensors_close(output[:, :3, :3], expected_slice, atol=1e-3)


@require_torch
Expand Down