EngineX/xc-llm-ascend
3
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

upgrade vLLM to main (#4608)

Browse Source 
1. fix https://github.com/vllm-project/vllm/pull/28542
The model structure modifications we involved in are:
     - Qwen2.5-VL(still exist some patch)
     - Qwen2-VL
     - Qwen2
     - DeepSeek series
     - Qwen-moe series
2. fix https://github.com/vllm-project/vllm/pull/29121
   the output token now  type changed from np to `list[list[int]]`

3. fix https://github.com/vllm-project/vllm/pull/29262
    `xformers` backend for multimodal now has been deprecated
4. fix https://github.com/vllm-project/vllm/pull/29342

5. fix https://github.com/vllm-project/vllm/pull/28579
6. fix https://github.com/vllm-project/vllm/pull/28718
7. fix https://github.com/vllm-project/vllm/issues/28665
8. fix https://github.com/vllm-project/vllm/pull/26847
vllm introduced the `optimization-level`, some default config has been
changed, and the param `--enforce-eager` has been deprecated
9. fix http://github.com/vllm-project/vllm/pull/29223 it retuns tuple
for sampler.
10. fix https://github.com/vllm-project/vllm/pull/29471 we'll remove the
related patch to avoid this kind of error.

Co-authored-by: hfadzxy <starmoon_zhang@163.com>
Co-authored-by: wangli <wangli858794774@gmail.com>


- vLLM version: v0.11.2

---------

Signed-off-by: wangxiyuan <wangxiyuan1007@gmail.com>
Signed-off-by: wangli <wangli858794774@gmail.com>
Signed-off-by: hfadzxy <starmoon_zhang@163.com>
Co-authored-by: wangli <wangli858794774@gmail.com>
Co-authored-by: hfadzxy <starmoon_zhang@163.com>
This commit is contained in:
wangxiyuan
2025-12-02 22:10:52 +08:00
committed by GitHub
parent 4588cdac02
commit 7f2673ea2d
60 changed files with 383 additions and 374 deletions
Download Patch File Download Diff File Expand all files Collapse all files

45
tests/ut/attention/test_mla_v1.py
View File

@@ -191,7 +191,7 @@ class TestAscendMLAMetadataBuilder(TestBase):
mock_vllm_config.cache_config.block_size = 16
mock_vllm_config.scheduler_config.max_num_seqs = 4
mock_vllm_config.scheduler_config.decode_max_num_seqs = 4
mock_vllm_config.scheduler_config.chunked_prefill_enabled = False
mock_vllm_config.scheduler_config.enable_chunked_prefill = False
mock_device = 'cpu'
mock_dcp.world_size = 1
@@ -213,7 +213,7 @@ class TestAscendMLAMetadataBuilder(TestBase):
mock_vllm_config.cache_config.block_size)
self.assertEqual(
builder.chunked_prefill_enabled,
mock_vllm_config.scheduler_config.chunked_prefill_enabled)
mock_vllm_config.scheduler_config.enable_chunked_prefill)
@patch('vllm.distributed.parallel_state.get_dcp_group')
@patch('vllm.distributed.parallel_state._DCP',
@@ -230,7 +230,7 @@ class TestAscendMLAMetadataBuilder(TestBase):
mock_vllm_config.cache_config.block_size = 16
mock_vllm_config.scheduler_config.max_num_seqs = 4
mock_vllm_config.scheduler_config.decode_max_num_seqs = 4
mock_vllm_config.scheduler_config.chunked_prefill_enabled = False
mock_vllm_config.scheduler_config.enable_chunked_prefill = False
mock_device = 'cpu'
mock_dcp.world_size = 1
@@ -254,7 +254,7 @@ class TestAscendMLAMetadataBuilder(TestBase):
mock_vllm_config.cache_config.block_size)
self.assertEqual(
builder.chunked_prefill_enabled,
mock_vllm_config.scheduler_config.chunked_prefill_enabled)
mock_vllm_config.scheduler_config.enable_chunked_prefill)
@patch('vllm.distributed.parallel_state.get_dcp_group')
@patch('vllm.distributed.parallel_state._DCP',
@@ -321,7 +321,7 @@ class TestAscendMLAMetadataBuilder(TestBase):
mock_vllm_config.cache_config.block_size = 16
mock_vllm_config.scheduler_config.max_num_seqs = 4
mock_vllm_config.scheduler_config.decode_max_num_seqs = 4
mock_vllm_config.scheduler_config.chunked_prefill_enabled = False
mock_vllm_config.scheduler_config.enable_chunked_prefill = False
mock_device = 'cpu'
mock_dcp.world_size = 1
@@ -440,8 +440,10 @@ class TestAscendMLAMetadataBuilderBuild(TestBase):
self.mock_vllm_config.model_config = ModelConfig(max_model_len=2048)
self.mock_vllm_config.model_config.hf_text_config.qk_rope_head_dim = 32
self.mock_vllm_config.cache_config = CacheConfig(block_size=32)
self.mock_vllm_config.scheduler_config = SchedulerConfig(
max_num_seqs=8, chunked_prefill_enabled=True)
mock_scheduler_config = MagicMock(spec=SchedulerConfig)
mock_scheduler_config.max_num_seqs = 8
mock_scheduler_config.chunked_prefill_enabled = True
self.mock_vllm_config.scheduler_config = mock_scheduler_config
self.mock_vllm_config.speculative_config = None
self.mock_device = torch.device("cpu")
@@ -454,12 +456,20 @@ class TestAscendMLAMetadataBuilderBuild(TestBase):
"vllm_ascend.attention.mla_v1.get_decode_context_model_parallel_world_size"
)
@patch("vllm_ascend.attention.mla_v1.get_ascend_config")
def test_build_prefix_no_cache_metadata(self, mock_get_ascend_config,
@patch("vllm_ascend.attention.mla_v1.torch.zeros", wraps=torch.zeros)
@patch("torch.Tensor.npu", new=lambda self: self)
@patch("torch.npu.is_available")
def test_build_prefix_no_cache_metadata(self, mock_npu_available,
mock_zeros, mock_get_ascend_config,
mock_dcp_world_size):
if not torch.npu.is_available():
self.skipTest("NPU not available, skipping NPU-dependent tests")
mock_npu_available.return_value = False
mock_dcp_world_size.return_value = 1
def zeros_override(*args, **kwargs):
kwargs.pop('pin_memory', None)
return mock_zeros._mock_wraps(*args, **kwargs)
mock_zeros.side_effect = zeros_override
common_attn_metadata = AscendCommonAttentionMetadata(
query_start_loc=torch.tensor([0, 3, 7]),
query_start_loc_cpu=torch.tensor([0, 3, 7]),
@@ -506,12 +516,21 @@ class TestAscendMLAMetadataBuilderBuild(TestBase):
"vllm_ascend.attention.mla_v1.get_decode_context_model_parallel_world_size"
)
@patch("vllm_ascend.attention.mla_v1.get_ascend_config")
def test_build_chunked_prefix_metadata(self, mock_get_ascend_config,
@patch("vllm_ascend.attention.mla_v1.torch.zeros", wraps=torch.zeros)
@patch("torch.Tensor.npu", new=lambda self: self)
@patch("torch.npu.is_available")
def test_build_chunked_prefix_metadata(self, mock_npu_available,
mock_zeros, mock_get_ascend_config,
mock_dcp_world_size):
if not torch.npu.is_available():
self.skipTest("NPU not available, skipping NPU-dependent tests")
mock_npu_available.return_value = False
mock_dcp_world_size.return_value = 1
def zeros_override(*args, **kwargs):
kwargs.pop('pin_memory', None)
return mock_zeros._mock_wraps(*args, **kwargs)
mock_zeros.side_effect = zeros_override
common_attn_metadata = AscendCommonAttentionMetadata(
query_start_loc=torch.tensor([0, 2, 5, 9]),
query_start_loc_cpu=torch.tensor([0, 2, 5, 9]),