SUMIN/HuggingFace_transformer
1
0
Fork 0
Code Issues Pull Requests Actions 7 Packages Projects Releases Wiki Activity

🚨🚨🚨 Enforce single model initialization (#21431)

Browse Source 
* Enforce single model initialization

* Add OneFormer example for problem 3

* Do it the Stas way

* Actually rename the uses...

* Rewrite test

* Try to change the test this way

* Fix all init slow/fast tests

* Break connection

* Fix more tests

* Fix test for initialization

* Remove custom test

* Quality

* Fix last failing tests

* The end?
This commit is contained in:
Sylvain Gugger
2023-02-09 15:46:26 -05:00
committed by GitHub
parent 2020ac4bd6
commit 04b2f13c37
25 changed files with 277 additions and 123 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
tests/models/prophetnet/test_modeling_prophetnet.py
View File

@@ -436,10 +436,10 @@ class ProphetNetModelTester:
decoder_attention_mask=decoder_attention_mask,
labels=lm_labels,
)
self.parent.assertTrue(torch.allclose(result.loss, torch.tensor(4.5819, device=torch_device), atol=1e-3))
self.parent.assertTrue(torch.allclose(result.loss, torch.tensor(4.5981, device=torch_device), atol=1e-3))
expected_logit_slice = torch.tensor(
[-0.1565, 0.0418, 0.1207, 0.0030, 0.0665, 0.0467, 0.0412], device=torch_device
[-0.0648, 0.0790, 0.0360, 0.0089, 0.0039, -0.0639, 0.0131], device=torch_device
)
self.parent.assertTrue(torch.allclose(result.logits[0, :, 1], expected_logit_slice, atol=1e-3))