File size: 10,648 Bytes

---

license: mit
base_model:
- OpenGVLab/InternViT-300M-448px
- internlm/internlm2_5-7b-chat
base_model_relation: merge
language:
- multilingual
pipeline_tag: image-text-to-text
library_name: transformers
tags:
- internvl
- video
- token compression
---


# PVC-InternVL2-8B

[\[📜 Paper\]](https://arxiv.org/abs/2412.09613)
[\[📂 GitHub\]](https://github.com/OpenGVLab/PVC)
[\[🚀 Quick Start\]](#quick-start)

## Introduction

We introduce the **Progressive Visual Token Compression (PVC)** in large vision-language models (VLMs), which unifies the visual inputs as videos and progressively compresses vision tokens across video frames. Our PVC achieves:

* Preserve spatial details and temporal dynamics for both images and videos.
* Effectively reduce the tokens used for each video frame and image tile.
* SoTA performance on various video benchmarks, including long and fine-grained short video tasks.
* No performance loss on image benchmarks, especially on detail-sensitive tasks.

<div style="text-align: center;">
    <img src="./assets/overview.png" width="70%"/>

</div>


## Results

Our implementation is based on the [InternVL2](https://github.com/OpenGVLab/InternVL) model, referred to as **PVC<sub>InternVL2</sub>**

### Video Understanding Benckmarks

| Model | LLaVA-OneVision-7B | Qwen2-VL-7B | InternVL2-8B | PVC<sub>InternVL2</sub>-8B |
| :--------------: | :--: | :--: | :--: | :--: |
| \# token/frame   | 196  | -    | 256  | 64   |
|                  |      |      |      |      |
| MVbench          | 56.7 | 67.0 | 66.4 | 73.8 |
| VideoMME w/o-sub | 58.2 | 63.3 | 54.0 | 64.1 |
| VideoMME w-sub   | 61.5 | 69.0 | 56.9 | 69.7 |
| MLVU             | 64.7 | -    | 52.0 | 72.4 |
| LongVideoBench   | 56.5 | -    | -    | 59.2 |
| NextQA           | 79.4 | -    | -    | 82.0 |
| Egoschema        | 60.1 | 66.7 | 55.0 | 59.6 |
| PercepTest       | 57.1 | 62.3 | 52.0 | 68.4 |
| AcNet-QA         | 56.6 | -    | -    | 57.1 |

### Image Understanding Benckmarks

| Model | LLaVA-OneVision-7B | Qwen2-VL-7B | InternVL2-8B | PVC<sub>InternVL2</sub>-8B |
| :--------------------: | :--: | :--: | :--: | :--: |
| \# token/image tile    | 729  | -    | 256  | 64   |
|                        |      |      |      |      |
| AI2D<sub>test</sub>    | 81.4 | 83.0 | 83.8 | 83.8 |
| ChartQA<sub>test</sub> | 80.0 | 83.0 | 83.3 | 84.1 |
| DocVQA<sub>test</sub>  | 87.5 | 94.5 | 91.6 | 92.5 |
| InfoVQA<sub>test</sub> | 68.8 | 76.5 | 74.8 | 75.0 |
| SQA<sub>test</sub>     | 96.0 | -    | 97.1 | 97.7 |
| TextVQA<sub>val</sub>  | -    | 84.3 | 77.4 | 80.0 |
| MMB<sub>en-test</sub>  | -    | 83.0 | 81.7 | 83.9 |
| MME<sub>sum</sub>      | 1998 | 2327 | 2210 | 2282 |
| MMMU<sub>val</sub>     | 48.8 | 54.1 | 49.3 | 50.9 |
| SEED<sub>I</sub>       | 75.4 | -    | 76.2 | 77.2 |
| OCRBench               | -    | 866  | 794  | 807  |

## Quick Start

```python

import numpy as np

import torch

import torchvision.transforms as T

from decord import VideoReader, cpu

from PIL import Image

from torchvision.transforms.functional import InterpolationMode

from transformers import AutoModel, AutoTokenizer



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



def get_index(bound, fps, max_frame, first_idx=0, num_segments=32):

    if bound:

        start, end = bound[0], bound[1]

    else:

        start, end = -100000, 100000

    start_idx = max(first_idx, round(start * fps))

    end_idx = min(round(end * fps), max_frame)

    seg_size = float(end_idx - start_idx) / num_segments

    frame_indices = np.array([

        int(start_idx + (seg_size / 2) + np.round(seg_size * idx))

        for idx in range(num_segments)

    ])

    return frame_indices



def load_video(video_path, bound=None, input_size=448, max_num=1, num_segments=32):

    vr = VideoReader(video_path, ctx=cpu(0), num_threads=1)

    max_frame = len(vr) - 1

    fps = float(vr.get_avg_fps())



    pixel_values_list, num_patches_list = [], []

    transform = build_transform(input_size=input_size)

    frame_indices = get_index(bound, fps, max_frame, first_idx=0, num_segments=num_segments)

    for frame_index in frame_indices:

        img = Image.fromarray(vr[frame_index].asnumpy()).convert('RGB')

        img = dynamic_preprocess(img, image_size=input_size, use_thumbnail=True, max_num=max_num)

        pixel_values = [transform(tile) for tile in img]

        pixel_values = torch.stack(pixel_values)

        num_patches_list.append(pixel_values.shape[0])

        pixel_values_list.append(pixel_values)

    pixel_values = torch.cat(pixel_values_list)

    return pixel_values, num_patches_list





path = 'OpenGVLab/PVC-InternVL2-8B'

model = AutoModel.from_pretrained(

    path,

    torch_dtype=torch.bfloat16,

    low_cpu_mem_usage=True,

    trust_remote_code=True).eval().cuda()

tokenizer = AutoTokenizer.from_pretrained(path, trust_remote_code=True, use_fast=False)

generation_config = dict(max_new_tokens=1024, do_sample=True)



# single-image conversation

pixel_values = load_image('./assets/example_image1.jpg', max_num=12).to(torch.bfloat16).cuda()

data_flag = torch.tensor([1], dtype=torch.long).cuda()



question = '<image>\nWhat is in the image?'

response = model.chat(tokenizer, pixel_values, question, generation_config, data_flag=data_flag)

print(f'User: {question}\nAssistant: {response}')



# multi-image conversation

pixel_values1 = load_image('./assets/example_image1.jpg', max_num=12).to(torch.bfloat16).cuda()

pixel_values2 = load_image('./assets/example_image2.jpg', max_num=12).to(torch.bfloat16).cuda()

pixel_values = torch.cat((pixel_values1, pixel_values2), dim=0)

data_flag = torch.tensor([2], dtype=torch.long).cuda()

num_patches_list = [pixel_values1.shape[0], pixel_values2.shape[0]]



question = 'Image-1: <image>\nImage-2: <image>\nWhat are the similarities and differences between these two images.'

response = model.chat(tokenizer, pixel_values, question, generation_config, data_flag=data_flag, num_patches_list=num_patches_list)

print(f'User: {question}\nAssistant: {response}')



# video conversation

pixel_values, num_patches_list = load_video('./assets/example_video.mp4', num_segments=64, max_num=1)

pixel_values = pixel_values.to(torch.bfloat16).cuda()

video_prefix = ''.join([f'Frame{i+1}: <image>\n' for i in range(len(num_patches_list))])

# Frame1: <image>\nFrame2: <image>\n...\nFrameN: <image>\n{question}

data_flag = torch.tensor([3], dtype=torch.long).cuda()



question = video_prefix + 'Describe this video in detail.'

response = model.chat(tokenizer, pixel_values, question, generation_config, data_flag=data_flag, num_patches_list=num_patches_list)

print(f'User: {question}\nAssistant: {response}')

```

## Evaluation

Please refer to our [Github Repo](https://github.com/OpenGVLab/PVC?tab=readme-ov-file#-evaluation).

## Citation

If you find this work helpful in your research, please consider citing:

```bibtex

@article{yang2024pvc,

  title={PVC: Progressive Visual Token Compression for Unified Image and Video Processing in Large Vision-Language Models},

  author={Yang, Chenyu and Dong, Xuan and Zhu, Xizhou and Su, Weijie and Wang, Jiahao and Tian, Hao and Chen, Zhe and Wang, Wenhai and Lu, Lewei and and Dai, Jifeng},

  journal={arXiv preprint arXiv:2412.09613},

  year={2024}

}

```

## License

This project is released under the MIT license. Parts of this project contain code and models from other sources, which are subject to their respective licenses.