Support nvidia/NVIDIA-Nemotron-Nano-12B-v2-VL-BF16 (and nvidia/C-RADIOv2-H) (#12277)

benchmark/mmmu/bench_hf.py

@@ -36,9 +36,10 @@ def eval_mmmu(args):
         try:
             # check if the model is belongs to internvl
             if "InternVL" in args.model_path:
-                from internvl_utils import load_image
                 from transformers import AutoTokenizer
 
+                from sglang.srt.multimodal.internvl_utils import image_to_pixel_values
+
                 tokenizer = AutoTokenizer.from_pretrained(args.model_path)
                 model = AutoModel.from_pretrained(
                     args.model_path,
@@ -80,7 +81,11 @@ def eval_mmmu(args):
         assert image is not None
 
         if "InternVL" in args.model_path:
-            pixel_values = load_image(sample["image_path"]).to(torch.bfloat16).cuda()
+            image = PIL.Image.open(sample["image_path"]).convert("RGB")
+            pixel_values = image_to_pixel_values(
+                image, input_size=448, max_num=12, use_thumbnail=True
+            )
+            pixel_values = pixel_values.to(device="cuda", dtype=torch.bfloat16)
             contents = ""
             if prefix:
                 contents += prefix

benchmark/mmmu/internvl_utils.py

@@ -1,94 +0,0 @@
-# copy from https://huggingface.co/OpenGVLab/InternVL3-1B
-import torch
-import torchvision.transforms as T
-from PIL import Image
-from torchvision.transforms.functional import InterpolationMode
-
-IMAGENET_MEAN = (0.485, 0.456, 0.406)
-IMAGENET_STD = (0.229, 0.224, 0.225)
-
-
-def build_transform(input_size):
-    MEAN, STD = IMAGENET_MEAN, IMAGENET_STD
-    transform = T.Compose(
-        [
-            T.Lambda(lambda img: img.convert("RGB") if img.mode != "RGB" else img),
-            T.Resize((input_size, input_size), interpolation=InterpolationMode.BICUBIC),
-            T.ToTensor(),
-            T.Normalize(mean=MEAN, std=STD),
-        ]
-    )
-    return transform
-
-
-def find_closest_aspect_ratio(aspect_ratio, target_ratios, width, height, image_size):
-    best_ratio_diff = float("inf")
-    best_ratio = (1, 1)
-    area = width * height
-    for ratio in target_ratios:
-        target_aspect_ratio = ratio[0] / ratio[1]
-        ratio_diff = abs(aspect_ratio - target_aspect_ratio)
-        if ratio_diff < best_ratio_diff:
-            best_ratio_diff = ratio_diff
-            best_ratio = ratio
-        elif ratio_diff == best_ratio_diff:
-            if area > 0.5 * image_size * image_size * ratio[0] * ratio[1]:
-                best_ratio = ratio
-    return best_ratio
-
-
-def dynamic_preprocess(
-    image, min_num=1, max_num=12, image_size=448, use_thumbnail=False
-):
-    orig_width, orig_height = image.size
-    aspect_ratio = orig_width / orig_height
-
-    # calculate the existing image aspect ratio
-    target_ratios = set(
-        (i, j)
-        for n in range(min_num, max_num + 1)
-        for i in range(1, n + 1)
-        for j in range(1, n + 1)
-        if i * j <= max_num and i * j >= min_num
-    )
-    target_ratios = sorted(target_ratios, key=lambda x: x[0] * x[1])
-
-    # find the closest aspect ratio to the target
-    target_aspect_ratio = find_closest_aspect_ratio(
-        aspect_ratio, target_ratios, orig_width, orig_height, image_size
-    )
-
-    # calculate the target width and height
-    target_width = image_size * target_aspect_ratio[0]
-    target_height = image_size * target_aspect_ratio[1]
-    blocks = target_aspect_ratio[0] * target_aspect_ratio[1]
-
-    # resize the image
-    resized_img = image.resize((target_width, target_height))
-    processed_images = []
-    for i in range(blocks):
-        box = (
-            (i % (target_width // image_size)) * image_size,
-            (i // (target_width // image_size)) * image_size,
-            ((i % (target_width // image_size)) + 1) * image_size,
-            ((i // (target_width // image_size)) + 1) * image_size,
-        )
-        # split the image
-        split_img = resized_img.crop(box)
-        processed_images.append(split_img)
-    assert len(processed_images) == blocks
-    if use_thumbnail and len(processed_images) != 1:
-        thumbnail_img = image.resize((image_size, image_size))
-        processed_images.append(thumbnail_img)
-    return processed_images
-
-
-def load_image(image_file, input_size=448, max_num=12):
-    image = Image.open(image_file).convert("RGB")
-    transform = build_transform(input_size=input_size)
-    images = dynamic_preprocess(
-        image, image_size=input_size, use_thumbnail=True, max_num=max_num
-    )
-    pixel_values = [transform(image) for image in images]
-    pixel_values = torch.stack(pixel_values)
-    return pixel_values