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

Cache: don't throw warnings on gemma2 when instantiating a new cache (#33595)

Browse Source
This commit is contained in:
Joao Gante
2024-09-19 17:42:47 +01:00
committed by GitHub
parent b50ff5993a
commit 52920b5dd5
4 changed files with 38 additions and 30 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
tests/models/gemma2/test_modeling_gemma2.py
View File

@@ -86,10 +86,15 @@ class Gemma2ModelTest(GemmaModelTest, unittest.TestCase):
def test_model_outputs_equivalence(self, **kwargs):
pass
@parameterized.expand([("float16",), ("bfloat16",), ("float32",)])
@unittest.skip("Gemma2's eager attn/sdpa attn outputs are expected to be different")
def test_eager_matches_sdpa_inference(self):
pass
@unittest.skip("Gemma2's eager attn/sdpa attn outputs are expected to be different")
def test_eager_matches_sdpa_generate(self):
pass
@parameterized.expand([("random",), ("same",)])
@unittest.skip("Gemma2 has HybridCache which is not compatible with assisted decoding")
def test_assisted_decoding_matches_greedy_search(self, assistant_type):