[Misc] Move some multimodal utils to modality-specific modules (#11494)

Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>

tests/models/decoder_only/vision_language/test_awq.py

@@ -3,7 +3,7 @@ from typing import List, Optional, Type
 import pytest
 import torch
 
-from vllm.multimodal.utils import rescale_image_size
+from vllm.multimodal.image import rescale_image_size
 
 from ....conftest import IMAGE_ASSETS, VllmRunner, _ImageAssets
 from ...utils import check_logprobs_close

tests/models/decoder_only/vision_language/test_h2ovl.py

@@ -8,7 +8,7 @@ from transformers import AutoConfig
 # Import the functions to test
 from vllm.model_executor.models.h2ovl import (calculate_num_blocks,
                                               image_to_pixel_values_wrapper)
-from vllm.multimodal.utils import rescale_image_size
+from vllm.multimodal.image import rescale_image_size
 
 models = [
     "h2oai/h2ovl-mississippi-800m",  # Replace with your actual model names

tests/models/decoder_only/vision_language/test_phi3v.py

@@ -5,7 +5,7 @@ from typing import List, Optional, Tuple, Type
 import pytest
 from transformers import AutoTokenizer
 
-from vllm.multimodal.utils import rescale_image_size
+from vllm.multimodal.image import rescale_image_size
 from vllm.platforms import current_platform
 from vllm.sequence import SampleLogprobs
 

tests/models/decoder_only/vision_language/test_qwen2_vl.py

@@ -6,8 +6,8 @@ import torch
 from PIL import Image
 
 from vllm.entrypoints.llm import LLM
-from vllm.multimodal.utils import (rescale_image_size, rescale_video_size,
-                                   sample_frames_from_video)
+from vllm.multimodal.image import rescale_image_size
+from vllm.multimodal.video import rescale_video_size, sample_frames_from_video
 
 from ....conftest import (IMAGE_ASSETS, VIDEO_ASSETS, PromptImageInput,
                           PromptVideoInput, VllmRunner)

tests/models/decoder_only/vision_language/vlm_utils/builders.py

@@ -5,8 +5,9 @@ from typing import Callable, Iterable, List, Optional, Tuple, Union
 
 import torch
 
-from vllm.multimodal.utils import (rescale_image_size, rescale_video_size,
-                                   resize_video, sample_frames_from_video)
+from vllm.multimodal.image import rescale_image_size
+from vllm.multimodal.video import (rescale_video_size, resize_video,
+                                   sample_frames_from_video)
 
 from .....conftest import _ImageAssets, _VideoAssets
 from .types import (SINGLE_IMAGE_BASE_PROMPTS, TEST_IMG_PLACEHOLDER,

tests/models/decoder_only/vision_language/vlm_utils/custom_inputs.py

@@ -1,8 +1,9 @@
 """Custom input builders for edge-cases in different models."""
 from typing import Callable
 
-from vllm.multimodal.utils import (rescale_image_size, rescale_video_size,
-                                   resize_video, sample_frames_from_video)
+from vllm.multimodal.image import rescale_image_size
+from vllm.multimodal.video import (rescale_video_size, resize_video,
+                                   sample_frames_from_video)
 
 from .....conftest import IMAGE_ASSETS, VIDEO_ASSETS
 from .builders import build_multi_image_inputs, build_single_image_inputs