Skip to content

[MM][Model][Perf] Remove Qwen2.5-VL modeling files and add patch for VisionAttention #4349

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 16 commits into from
Nov 28, 2025
Merged

[MM][Model][Perf] Remove Qwen2.5-VL modeling files and add patch for VisionAttention #4349

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
16 commits
Select commit Hold shift + click to select a range
d4c8b87
patch AscendQwen2_5_VisionAttention
shen-shanshan Nov 21, 2025
29fae87
remove more modeling files
shen-shanshan Nov 24, 2025
a4b291a
remove qwen2.5-vl ut
shen-shanshan Nov 24, 2025
c198058
add padding manager
shen-shanshan Nov 24, 2025
a674f80
sync main
shen-shanshan Nov 24, 2025
f391ac1
remove qwen2.5-vl without padding
shen-shanshan Nov 25, 2025
73ab6c4
update
shen-shanshan Nov 25, 2025
26b12d6
add get_forward_context_manager interface
shen-shanshan Nov 25, 2025
0310c88
modify pad logic
shen-shanshan Nov 26, 2025
a70e857
add patch for rope
shen-shanshan Nov 27, 2025
f0af5cc
fix lint
shen-shanshan Nov 27, 2025
f3a6ad8
fix lint
shen-shanshan Nov 27, 2025
e02e99c
fix lint
shen-shanshan Nov 27, 2025
523ab29
fix lint
shen-shanshan Nov 27, 2025
f27ac29
fix lint
shen-shanshan Nov 27, 2025
9c62d27
fix platform
shen-shanshan Nov 28, 2025
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
488 changes: 0 additions & 488 deletions tests/ut/models/test_qwen2_5_vl.py
View file
Open in desktop

This file was deleted.

422 changes: 0 additions & 422 deletions tests/ut/models/test_qwen2_5_vl_without_padding.py
View file
Open in desktop

This file was deleted.

19 changes: 2 additions & 17 deletions vllm_ascend/models/__init__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,7 +1,5 @@
from vllm import ModelRegistry

import vllm_ascend.envs as envs_ascend


def register_model():
ModelRegistry.register_model(
Expand All @@ -10,24 +8,11 @@ def register_model():

ModelRegistry.register_model(
"Qwen3VLMoeForConditionalGeneration",
"vllm_ascend.models.qwen2_5_vl_without_padding:AscendQwen3VLMoeForConditionalGeneration"
)
"vllm_ascend.models.qwen3_vl:AscendQwen3VLMoeForConditionalGeneration")

ModelRegistry.register_model(
"Qwen3VLForConditionalGeneration",
"vllm_ascend.models.qwen2_5_vl_without_padding:AscendQwen3VLForConditionalGeneration"
)

if envs_ascend.USE_OPTIMIZED_MODEL:
ModelRegistry.register_model(
"Qwen2_5_VLForConditionalGeneration",
"vllm_ascend.models.qwen2_5_vl:AscendQwen2_5_VLForConditionalGeneration"
)
else:
ModelRegistry.register_model(
"Qwen2_5_VLForConditionalGeneration",
"vllm_ascend.models.qwen2_5_vl_without_padding:AscendQwen2_5_VLForConditionalGeneration_Without_Padding"
)
"vllm_ascend.models.qwen3_vl:AscendQwen3VLForConditionalGeneration")

# There is no PanguProMoEForCausalLM in vLLM, so we should register it before vLLM config initialization
# to make sure the model can be loaded correctly. This register step can be removed once vLLM support PanguProMoEForCausalLM.
Expand Down
Loading
Loading