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

Fix CI for VLMs (#35690)

Browse Source 
* fix some easy test

* more tests

* remove logit check here also

* add require_torch_large_gpu in Emu3
This commit is contained in:
Raushan Turganbay
2025-01-20 11:15:39 +01:00
committed by GitHub
parent 5fa3534475
commit 8571bb145a
17 changed files with 102 additions and 485 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
tests/models/instructblip/test_modeling_instructblip.py
View File

@@ -842,22 +842,11 @@ class InstructBlipModelIntegrationTest(unittest.TestCase):
prompt = "What is unusual about this image?"
inputs = processor(images=image, text=prompt, return_tensors="pt").to(torch_device, torch.float16)
# verify logits
with torch.no_grad():
logits = model(**inputs).logits
expected_slice = torch.tensor(
[[-3.3047, -12.0625, 8.4922], [-4.9258, -11.7578, 8.1406], [-3.9297, -13.5000, 9.2500]],
device=torch_device,
)
self.assertTrue(torch.allclose(logits[0, :3, :3].float(), expected_slice, atol=1e-3))
# verify generation
outputs = model.generate(**inputs, max_new_tokens=30)
generated_text = processor.batch_decode(outputs, skip_special_tokens=True)[0].strip()
expected_outputs = [2, 1724, 338, 22910, 1048, 445, 1967, 29973, 450, 22910, 9565, 310, 445, 1967, 338, 393, 263, 767, 338, 13977, 292, 22095, 373, 278, 1250, 310, 263, 13328, 20134, 29963, 1550, 19500, 373, 263, 19587, 4272, 11952, 29889] # fmt: off
expected_outputs = [32001] * 32 + [2, 1724, 338, 22910, 1048, 445, 1967, 29973, 450, 22910, 9565, 310, 445, 1967, 338, 393, 263, 767, 338, 13977, 292, 22095, 373, 278, 1250, 310, 263, 13328, 20134, 29963, 1550, 19500, 373, 263, 19587, 4272, 11952, 29889] # fmt: off
self.assertEqual(outputs[0].tolist(), expected_outputs)
self.assertEqual(