[3/N] Initialize MM components in context managers (I-L) (#32650)

Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
This commit is contained in:
Cyrus Leung
2026-01-20 10:21:56 +00:00
committed by GitHub
parent 8be263c3fb
commit 7f1bcd18ff
10 changed files with 203 additions and 389 deletions
+24 -28
View File
@@ -312,12 +312,10 @@ class LlavaNextVideoForConditionalGeneration(nn.Module, SupportsMultiModal, Supp
@classmethod
def get_placeholder_str(cls, modality: str, i: int) -> str | None:
if modality.startswith("image"):
return "<image>"
if modality.startswith("video"):
return "<video>"
raise ValueError("Only image or video modality is supported")
raise ValueError("Only video modality is supported")
def __init__(self, *, vllm_config: VllmConfig, prefix: str = "") -> None:
super().__init__()
@@ -329,26 +327,29 @@ class LlavaNextVideoForConditionalGeneration(nn.Module, SupportsMultiModal, Supp
self.config = config
self.multimodal_config = multimodal_config
# Initialize the vision tower only up to the required feature layer
self.vision_tower = init_vision_tower_for_llava(
config,
quant_config=quant_config,
multimodal_config=multimodal_config,
require_post_norm=False,
prefix=maybe_prefix(prefix, "vision_tower"),
)
self.vision_resampler = LlavaNextVideoPooler(config)
self.multi_modal_projector = LlavaNextMultiModalProjector(
vision_hidden_size=config.vision_config.hidden_size,
text_hidden_size=config.text_config.hidden_size,
projector_hidden_act=config.projector_hidden_act,
multimodal_projector_bias=config.multimodal_projector_bias,
)
self.language_model = init_vllm_registered_model(
vllm_config=vllm_config,
hf_config=config.text_config,
prefix=maybe_prefix(prefix, "language_model"),
)
with self._mark_tower_model(vllm_config, "video"):
# Initialize the vision tower only up to the required feature layer
self.vision_tower = init_vision_tower_for_llava(
config,
quant_config=quant_config,
multimodal_config=multimodal_config,
require_post_norm=False,
prefix=maybe_prefix(prefix, "vision_tower"),
)
self.vision_resampler = LlavaNextVideoPooler(config)
self.multi_modal_projector = LlavaNextMultiModalProjector(
vision_hidden_size=config.vision_config.hidden_size,
text_hidden_size=config.text_config.hidden_size,
projector_hidden_act=config.projector_hidden_act,
multimodal_projector_bias=config.multimodal_projector_bias,
)
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"),
)
self.make_empty_intermediate_tensors = (
self.language_model.model.make_empty_intermediate_tensors
@@ -395,8 +396,6 @@ class LlavaNextVideoForConditionalGeneration(nn.Module, SupportsMultiModal, Supp
return image_features
def _process_video_pixels(self, inputs: LlavaNextVideoPixelInputs):
assert self.vision_tower is not None
video_pixels = inputs["pixel_values_videos"]
if isinstance(video_pixels, torch.Tensor):
@@ -419,9 +418,6 @@ class LlavaNextVideoForConditionalGeneration(nn.Module, SupportsMultiModal, Supp
return [e.flatten(0, 1) for e in embeds]
def get_language_model(self) -> torch.nn.Module:
return self.language_model
def embed_multimodal(self, **kwargs: object) -> MultiModalEmbeddings:
video_input = self._parse_and_validate_video_input(**kwargs)
if video_input is None: