Skip to content
Open
Changes from all commits
Commits
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
3 changes: 2 additions & 1 deletion tests/openvino/test_genai.py
Original file line number Diff line number Diff line change
Expand Up @@ -571,6 +571,7 @@ def test_compare_outputs(self, model_arch):
torch.testing.assert_close(transformers_output, genai_output, rtol=1e-2, atol=1e-3)


# NOTE: DFlash wasn't merge yet to an official release so skipping the tests for now
@pytest.mark.skipif(OPENVINO_DEVICE == "NPU", reason="Speculative decoding tests are not yet supported on NPU")
class LLMPipelineWithSpeculativeDecodingTestCase(unittest.TestCase):
GEN_KWARGS = {
Expand All @@ -581,7 +582,7 @@ class LLMPipelineWithSpeculativeDecodingTestCase(unittest.TestCase):
}
SPECULATIVE_DECODING_MODELS = [
(model_arch, model_pair, "Eagle3", None, "2026.0") for model_arch, model_pair in EAGLE3_MODELS.items()
] + [(model_arch, model_pair, "DFlash", "4.57", "2026.3") for model_arch, model_pair in DFLASH_MODELS.items()]
] + [(model_arch, model_pair, "DFlash", "4.57", "2026.999") for model_arch, model_pair in DFLASH_MODELS.items()]

@parameterized.expand(SPECULATIVE_DECODING_MODELS)
def test_compare_outputs(
Expand Down