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

[tests] enable bnb tests on xpu (#36233)

Browse Source 
* fix failed test

* fix device

* fix more device cases

* add more cases

* fix empty cache

* Update test_4bit.py

---------

Co-authored-by: Yih-Dar <2521628+ydshieh@users.noreply.github.com>
This commit is contained in:
Fanli Lin
2025-02-24 18:30:15 +08:00
committed by GitHub
parent 92c5ca9dd7
commit 4dbf17c17f
4 changed files with 27 additions and 21 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
tests/models/falcon/test_modeling_falcon.py
View File

@@ -591,12 +591,12 @@ class FalconLanguageGenerationTest(unittest.TestCase):
test_text = "A sequence: 1, 2" # should generate the rest of the sequence
unpadded_inputs = tokenizer([test_text], return_tensors="pt").to("cuda:0")
unpadded_inputs = tokenizer([test_text], return_tensors="pt").to(f"{torch_device}:0")
unpadded_gen_out = model.generate(**unpadded_inputs, max_new_tokens=20)
unpadded_gen_text = tokenizer.batch_decode(unpadded_gen_out, skip_special_tokens=True)
dummy_text = "This is a longer text " * 2 # forces left-padding on `test_text`
padded_inputs = tokenizer([test_text, dummy_text], return_tensors="pt", padding=True).to("cuda:0")
padded_inputs = tokenizer([test_text, dummy_text], return_tensors="pt", padding=True).to(f"{torch_device}:0")
padded_gen_out = model.generate(**padded_inputs, max_new_tokens=20)
padded_gen_text = tokenizer.batch_decode(padded_gen_out, skip_special_tokens=True)