EngineX-Ascend/enginex-ascend-910-vllm
10
0
Fork 1
Code Issues Pull Requests Actions Projects Releases Wiki Activity

init v0.11.0rc0

Browse Source
This commit is contained in:
luopingyi
2025-10-14 10:38:28 +08:00
parent 67afd0ea78
commit 66dc16f966
278 changed files with 28130 additions and 11708 deletions
Download Patch File Download Diff File Expand all files Collapse all files

42
tests/ut/attention/test_mla_v1.py
View File

@@ -186,10 +186,39 @@ class TestAscendMLAMetadataBuilder(TestBase):
mock_vllm_config.scheduler_config.chunked_prefill_enabled = False
mock_device = 'cpu'
mock_vllm_config.speculative_config = None
ascend_config = MagicMock()
with patch("vllm_ascend.attention.mla_v1.get_ascend_config",
return_value=ascend_config):
builder = AscendMLAMetadataBuilder(mock_vllm_config, mock_device)
builder = AscendMLAMetadataBuilder(None, None, mock_vllm_config,
mock_device)
self.assertEqual(builder.block_size,
mock_vllm_config.cache_config.block_size)
self.assertEqual(
builder.chunked_prefill_enabled,
mock_vllm_config.scheduler_config.chunked_prefill_enabled)
def test_ascend_mla_metadata_builder_spec_decode(self):
mock_vllm_config = MagicMock()
mock_vllm_config.model_config.max_model_len = 1024
mock_vllm_config.model_config.get_head_size.return_value = 64
mock_vllm_config.model_config.dtype = torch.float16
mock_vllm_config.cache_config.block_size = 16
mock_vllm_config.scheduler_config.max_num_seqs = 4
mock_vllm_config.scheduler_config.chunked_prefill_enabled = False
mock_device = 'cpu'
mock_spec_config = MagicMock()
mock_spec_config.num_speculative_tokens = 3
mock_vllm_config.speculative_config = mock_spec_config
ascend_config = MagicMock()
with patch("vllm_ascend.attention.mla_v1.get_ascend_config",
return_value=ascend_config):
builder = AscendMLAMetadataBuilder(None, None, mock_vllm_config,
mock_device)
self.assertEqual(builder.block_size,
mock_vllm_config.cache_config.block_size)
@@ -207,9 +236,12 @@ class TestAscendMLAMetadataBuilder(TestBase):
mock_vllm_config.scheduler_config.chunked_prefill_enabled = False
mock_device = 'cpu'
mock_vllm_config.speculative_config = None
with patch("vllm_ascend.attention.mla_v1.get_ascend_config",
return_value=ascend_config):
builder = AscendMLAMetadataBuilder(mock_vllm_config, mock_device)
builder = AscendMLAMetadataBuilder(None, None, mock_vllm_config,
mock_device)
builder.decode_threshold = 1
input_batch = MagicMock()
@@ -522,7 +554,11 @@ class TestAscendMLAImpl(TestBase):
self.impl.num_kv_heads = self.impl.num_heads
decode_res, prefill_res = self.impl._mla_preprocess(
hidden_states, kv_cache, attn_metadata, need_gather_q_kv=False)
"mock_layer",
hidden_states,
kv_cache,
attn_metadata,
need_gather_q_kv=False)
self.assertIsNotNone(decode_res)
self.assertIsNotNone(prefill_res)