app_regression.py

import gradio as gr
import spaces
import os
import time
import json
import numpy as np
import av
import torch
from PIL import Image
import functools
from transformers import AutoProcessor, AutoConfig
from models.idefics2 import Idefics2ForSequenceClassification
from models.conversation import conv_templates
from typing import List


processor = AutoProcessor.from_pretrained("TIGER-Lab/VideoScore")
model = Idefics2ForSequenceClassification.from_pretrained("TIGER-Lab/VideoScore", torch_dtype=torch.bfloat16).eval()

MAX_NUM_FRAMES = 24
conv_template = conv_templates["idefics_2"]

with open("./examples/all_subsets.json", 'r') as f:
    examples = json.load(f)

for item in examples:
    video_id = item['images'][0].split("_")[0]
    item['images'] = [os.path.join("./examples", video_id, x) for x in item['images']]
    item['video'] = os.path.join("./examples", item['video'])
    
with open("./examples/hd.json", 'r') as f:
    hd_examples = json.load(f)

for item in hd_examples:
    item['video'] = os.path.join("./examples", item['video'])

examples = hd_examples + examples

VIDEO_EVAL_PROMPT = """
Suppose you are an expert in judging and evaluating the quality of AI-generated videos, 
please watch the following frames of a given video and see the text prompt for generating the video, 
then give scores from 5 different dimensions:
(1) visual quality: the quality of the video in terms of clearness, resolution, brightness, and color
(2) temporal consistency, the consistency of objects or humans in video
(3) dynamic degree, the degree of dynamic changes
(4) text-to-video alignment, the alignment between the text prompt and the video content
(5) factual consistency, the consistency of the video content with the common-sense and factual knowledge

For each dimension, output a number from [1,2,3,4], 
in which '1' means 'Bad', '2' means 'Average', '3' means 'Good', 
'4' means 'Real' or 'Perfect' (the video is like a real video)
Here is an output example:
visual quality: 4
temporal consistency: 4
dynamic degree: 3
text-to-video alignment: 1
factual consistency: 2

For this video, the text prompt is "{text_prompt}",
all the frames of video are as follows: 

"""



space_description="""\
[📃Paper](https://arxiv.org/abs/2406.15252) | [🌐Website](https://tiger-ai-lab.github.io/VideoScore/) | [💻Github](https://github.com/TIGER-AI-Lab/VideoScore) | [🛢️Datasets](https://huggingface.co/datasets/TIGER-Lab/VideoFeedback) | [🤗Model](https://huggingface.co/TIGER-Lab/VideoScore) | [🤗Demo](https://huggingface.co/spaces/TIGER-Lab/VideoScore)

- VideoScore is a video quality evaluation model, taking [Mantis-8B-Idefics2](https://huggingface.co/TIGER-Lab/Mantis-8B-Idefics2) as base-model
and trained on [VideoFeedback](https://huggingface.co/datasets/TIGER-Lab/VideoFeedback),
a large video evaluation dataset with multi-aspect human scores. 

- VideoScore can reach 75+ Spearman correlation with humans on VideoEval-test, surpassing all the MLLM-prompting methods and feature-based metrics. 

- VideoScore also beat the best baselines on other three benchmarks EvalCrafter, GenAI-Bench and VBench, showing high alignment with human evaluations.
"""


aspect_mapping= [
    "visual quality",
    "temporal consistency",
    "dynamic degree",
    "text-to-video alignment",
    "factual consistency",
]


@spaces.GPU(duration=60)
def score(prompt:str, images:List[Image.Image]):
    if not prompt:
        raise gr.Error("Please provide a prompt")
    model.to("cuda")
    if not images:
        images = None
    
    flatten_images = []
    for x in images:
        if isinstance(x, list):
            flatten_images.extend(x)
        else:
            flatten_images.append(x)
    
    flatten_images = [Image.open(x) if isinstance(x, str) else x for x in flatten_images]
    inputs = processor(text=prompt, images=flatten_images, return_tensors="pt")
    inputs = {k: v.to(model.device) for k, v in inputs.items()}
    with torch.no_grad():
        outputs = model(**inputs)
    
    logits = outputs.logits
    num_aspects = logits.shape[-1]
    aspects = [aspect_mapping[i] for i in range(num_aspects)]
    
    aspect_scores = {}
    for i, aspect in enumerate(aspects):
        aspect_scores[aspect] = round(logits[0, i].item(), 2)
    return aspect_scores
    
    
def read_video_pyav(container, indices):
    '''
    Decode the video with PyAV decoder.

    Args:
        container (av.container.input.InputContainer): PyAV container.
        indices (List[int]): List of frame indices to decode.

    Returns:
        np.ndarray: np array of decoded frames of shape (num_frames, height, width, 3).
    '''
    frames = []
    container.seek(0)
    start_index = indices[0]
    end_index = indices[-1]
    for i, frame in enumerate(container.decode(video=0)):
        if i > end_index:
            break
        if i >= start_index and i in indices:
            frames.append(frame)
    return np.stack([x.to_ndarray(format="rgb24") for x in frames])

def eval_video(prompt, video:str):
    container = av.open(video)

    # sample uniformly 8 frames from the video
    total_frames = container.streams.video[0].frames
    if total_frames > MAX_NUM_FRAMES:
        indices = np.arange(0, total_frames, total_frames / MAX_NUM_FRAMES).astype(int)
    else:
        indices = np.arange(total_frames)
    video_frames = read_video_pyav(container, indices)

    frames = [Image.fromarray(x) for x in video_frames]
    
    eval_prompt = VIDEO_EVAL_PROMPT.format(text_prompt=prompt)
    
    num_image_token = eval_prompt.count("<image>")
    if num_image_token < len(frames):
        eval_prompt += "<image> " * (len(frames) - num_image_token)
    
    aspect_scores = score(eval_prompt, [frames])
    return aspect_scores

def build_demo():
    with gr.Blocks() as demo:

        gr.Markdown("## VideoScore: Building Automatic Metrics to Simulate Fine-grained Human Feedback for Video Generation")
        with gr.Row():
            gr.Markdown(space_description)
            gr.Image("https://tiger-ai-lab.github.io/VideoScore/static/images/teaser.png", label="Teaser")
        
        gr.Markdown("### Try VideoScore (Regression) with your own text prompt and videos.")
        with gr.Row():
            video = gr.Video(width=500, label="Video")
            with gr.Column():
                eval_prompt_template = gr.Textbox(VIDEO_EVAL_PROMPT.strip(' \n'), label="Evaluation Prompt Template", interactive=False, max_lines=26)
                video_prompt = gr.Textbox(label="Text Prompt", lines=1)
                with gr.Row():
                    eval_button = gr.Button("Evaluate Video")
                    clear_button = gr.ClearButton([video, video_prompt])
                # eval_result = gr.Textbox(label="Evaluation result", interactive=False, lines=7)
                eval_result = gr.Json(label="Evaluation result")
                
        
        eval_button.click(
            eval_video, [video_prompt, video], [eval_result]
        )
        
        dummy_id = gr.Textbox("id", label="id", visible=False, min_width=50)
        # dummy_output = gr.Textbox("reference score", label="reference scores", visible=False, lines=7)
        
        gr.Examples(
            examples=
            [
                [
                    # item['id'], 
                    item['prompt'],
                    item['video'],
                    # item['conversations'][1]['value']
                ] for item in examples if item['prompt']
            ],
            inputs=[video_prompt, video],
            # inputs=[dummy_id, video_prompt, video, dummy_output],
            
        )        
        
        gr.Markdown("""
## Citation
```
@article{he2024videoscore,
  title = {VideoScore: Building Automatic Metrics to Simulate Fine-grained Human Feedback for Video Generation},
  author = {He, Xuan and Jiang, Dongfu and Zhang, Ge and Ku, Max and Soni, Achint and Siu, Sherman and Chen, Haonan and Chandra, Abhranil and Jiang, Ziyan and Arulraj, Aaran and Wang, Kai and Do, Quy Duc and Ni, Yuansheng and Lyu, Bohan and Narsupalli, Yaswanth and Fan, Rongqi and Lyu, Zhiheng and Lin, Yuchen and Chen, Wenhu},
  journal = {ArXiv},
  year = {2024},
  volume={abs/2406.15252},
  url = {https://arxiv.org/abs/2406.15252},
}
```""")
    return demo    
    

if __name__ == "__main__":
    demo = build_demo()
    demo.launch(share=True)