biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Model] Use context managers for encoder- and LM-only mode (#32605)

Browse Source 
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
This commit is contained in:
Cyrus Leung
2026-01-20 11:43:38 +08:00
committed by GitHub
parent 6c01ffb897
commit 4753f3bf69
21 changed files with 290 additions and 353 deletions
Download Patch File Download Diff File Expand all files Collapse all files

17
vllm/model_executor/models/opencua.py
View File

@@ -239,7 +239,7 @@ class OpenCUAForConditionalGeneration(Qwen2_5_VLForConditionalGeneration):
multimodal_config.is_multimodal_pruning_enabled()
)
if multimodal_config.get_limit_per_prompt("image"):
with self._mark_tower_model(vllm_config, "image"):
self.visual = OpenCUAVisionTransformer(
vision_config=config.vision_config,
norm_eps=getattr(config, "rms_norm_eps", 1e-6),
@@ -247,15 +247,14 @@ class OpenCUAForConditionalGeneration(Qwen2_5_VLForConditionalGeneration):
multimodal_config=self.multimodal_config,
prefix=maybe_prefix(prefix, "visual"),
)
else:
self.visual = None
self.language_model = init_vllm_registered_model(
vllm_config=vllm_config,
hf_config=config.text_config,
prefix=maybe_prefix(prefix, "language_model"),
architectures=["Qwen2ForCausalLM"],
)
with self._mark_language_model(vllm_config):
self.language_model = init_vllm_registered_model(
vllm_config=vllm_config,
hf_config=config.text_config,
prefix=maybe_prefix(prefix, "language_model"),
architectures=["Qwen2ForCausalLM"],
)
self.make_empty_intermediate_tensors = (
self.language_model.make_empty_intermediate_tensors