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

Implicit language-model-only mode via limit-mm-per-prompt (#22299)

Browse Source 
Signed-off-by: Roger Wang <hey@rogerw.me>
Signed-off-by: Andy Xie <andy.xning@gmail.com>
Signed-off-by: tjtanaa <tunjian.tan@embeddedllm.com>
Signed-off-by: Andrew Sansom <andrew@protopia.ai>
Signed-off-by: Zhiyu Cheng <zhiyuc@nvidia.com>
Signed-off-by: Shu Wang <shuw@nvidia.com>
Signed-off-by: Po-Han Huang <pohanh@nvidia.com>
Signed-off-by: Shu Wang. <shuw@nvidia.com>
Signed-off-by: XIn Li <xinli@nvidia.com>
Signed-off-by: Junhao Li <junhao@ubicloud.com>
Signed-off-by: chaunceyjiang <chaunceyjiang@gmail.com>
Signed-off-by: zRzRzRzRzRzRzR <2448370773@qq.com>
Signed-off-by: zitian.zhao <zitian.zhao@tencentmusic.com>
Signed-off-by: zitian zhao <zitian.zhao@tencentmusic.com>
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
Signed-off-by: iAmir97 <Amir.balwel@embeddedllm.com>
Signed-off-by: iAmir97 <71513472+iAmir97@users.noreply.github.com>
Signed-off-by: Linkun <github@lkchen.net>
Co-authored-by: Ning Xie <andy.xning@gmail.com>
Co-authored-by: TJian <tunjian.tan@embeddedllm.com>
Co-authored-by: Andrew Sansom <andrew@protopia.ai>
Co-authored-by: Zhiyu <zhiyuc@nvidia.com>
Co-authored-by: Shu Wang <shuw@nvidia.com>
Co-authored-by: XIn Li <xinli@nvidia.com>
Co-authored-by: Junhao Li <streaver91@gmail.com>
Co-authored-by: Chauncey <chaunceyjiang@gmail.com>
Co-authored-by: Yuxuan Zhang <2448370773@qq.com>
Co-authored-by: ZiTian Zhao <zitian.zhao@tencentmusic.com>
Co-authored-by: Cyrus Leung <tlleungac@connect.ust.hk>
Co-authored-by: Po-Han Huang (NVIDIA) <53919306+nvpohanh@users.noreply.github.com>
Co-authored-by: iAmir97 <71513472+iAmir97@users.noreply.github.com>
Co-authored-by: iAmir97 <Amir.balwel@embeddedllm.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: Cyrus Leung <cyrus.tl.leung@gmail.com>
Co-authored-by: Hong Hanh <hanh.usth@gmail.com>
Co-authored-by: youkaichao <youkaichao@gmail.com>
Co-authored-by: lkchen <github@lkchen.net>
This commit is contained in:
Roger Wang
2025-08-08 22:21:40 -07:00
committed by GitHub
parent 429e4e2d42
commit 08b751ba74
16 changed files with 271 additions and 116 deletions
Download Patch File Download Diff File Expand all files Collapse all files

60
vllm/model_executor/models/step3_vl.py
View File

@@ -837,27 +837,35 @@ class Step3VLForConditionalGeneration(nn.Module, SupportsMultiModal,
self.config = config
self.multimodal_config = multimodal_config
self.vision_model = Step3VisionTransformer(config.vision_config,
None,
prefix=maybe_prefix(
prefix, "vision_model"))
self.vit_downsampler = nn.Conv2d(
config.vision_config.hidden_size,
config.vision_config.output_hidden_size,
kernel_size=2,
stride=config.understand_projector_stride)
self.vit_downsampler2 = nn.Conv2d(
config.vision_config.output_hidden_size,
config.vision_config.output_hidden_size * 2,
kernel_size=3,
stride=2,
padding=1,
)
self.vit_large_projector = nn.Linear(
config.vision_config.output_hidden_size * 2,
config.hidden_size,
bias=config.projector_bias,
)
if multimodal_config.get_limit_per_prompt("image"):
self.vision_model = Step3VisionTransformer(config.vision_config,
None,
prefix=maybe_prefix(
prefix,
"vision_model"))
self.vit_downsampler = nn.Conv2d(
config.vision_config.hidden_size,
config.vision_config.output_hidden_size,
kernel_size=2,
stride=config.understand_projector_stride)
self.vit_downsampler2 = nn.Conv2d(
config.vision_config.output_hidden_size,
config.vision_config.output_hidden_size * 2,
kernel_size=3,
stride=2,
padding=1,
)
self.vit_large_projector = nn.Linear(
config.vision_config.output_hidden_size * 2,
config.hidden_size,
bias=config.projector_bias,
)
else:
self.vision_model = None
self.vit_downsampler = None
self.vit_downsampler2 = None
self.vit_large_projector = None
self.language_model = init_vllm_registered_model(
vllm_config=vllm_config,
hf_config=config.text_config,
@@ -1046,7 +1054,15 @@ class Step3VLForConditionalGeneration(nn.Module, SupportsMultiModal,
return self.language_model.sample(logits, sampling_metadata)
def load_weights(self, weights: Iterable[tuple[str, torch.Tensor]]):
loader = AutoWeightsLoader(self)
skip_prefixes = []
if self.vision_model is None and self.vit_large_projector is None:
skip_prefixes = [
"vision_model.", "vit_downsampler.", "vit_downsampler2.",
"vit_large_projector."
]
loader = AutoWeightsLoader(self, skip_prefixes=skip_prefixes)
loaded_weights = loader.load_weights(weights,
mapper=self.hf_to_vllm_mapper)
return loaded_weights