Improve model card: add pipeline_tag, library_name, link project page and update images

#4
by nielsr HF Staff - opened
Files changed (1) hide show
  1. README.md +17 -10
README.md CHANGED
@@ -1,10 +1,13 @@
1
  ---
2
- license: mit
3
- language:
4
- - en
5
  base_model:
6
  - Skywork/Matrix-Game
 
 
 
 
 
7
  ---
 
8
  <!-- markdownlint-disable first-line-h1 -->
9
  <!-- markdownlint-disable html -->
10
  <!-- markdownlint-disable no-duplicate-header -->
@@ -17,6 +20,9 @@ base_model:
17
  <a href="https://arxiv.org/pdf/2506.18701">
18
  <img src="https://img.shields.io/badge/arXiv-Report-b31b1b?style=flat&logo=arxiv&logoColor=white" alt="arXiv">
19
  </a>
 
 
 
20
 
21
 
22
  </div>
@@ -37,24 +43,25 @@ base_model:
37
  ## πŸš€ Performance Comparison
38
  ### GameWorld Score Benchmark Comparison
39
 
40
- | Model | Image Quality ↑ | Aesthetic Quality ↑ | Temporal Cons. ↑ | Motion Smooth. ↑ | Keyboard Acc. ↑ | Mouse Acc. ↑ | 3D Cons. ↑ |
41
- |-----------|------------------|-------------|-------------------|-------------------|------------------|---------------|-------------|
42
- | Oasis | 0.65 | 0.48 | 0.94 | **0.98** | 0.77 | 0.56 | 0.56 |
43
- | MineWorld | 0.69 | 0.47 | 0.95 | **0.98** | 0.86 | 0.64 | 0.51 |
44
- | **Ours** | **0.72** | **0.49** | **0.97** | **0.98** | **0.95** | **0.95** | **0.76** |
45
 
46
  **Metric Descriptions**:
47
 
48
  - **Image Quality** / **Aesthetic**: Visual fidelity and perceptual appeal of generated frames
49
  - **Temporal Consistency** / **Motion Smoothness**: Temporal coherence and smoothness between frames
50
  - **Keyboard Accuracy** / **Mouse Accuracy**: Accuracy in following user control signals
51
- - **3D Consistency**: Geometric stability and physical plausibility over time
 
52
 
53
  Please check our [GameWorld](https://github.com/SkyworkAI/Matrix-Game/tree/main/GameWorldScore) benchmark for detailed implementation.
54
 
55
  ### Human Evaluation
56
 
57
- ![Human Win Rate](assets/human_win_rate.png)
58
 
59
  > Double-blind human evaluation by two independent groups across four key dimensions: **Overall Quality**, **Controllability**, **Visual Quality**, and **Temporal Consistency**.
60
  > Scores represent the percentage of pairwise comparisons in which each method was preferred. Matrix-Game consistently outperforms prior models across all metrics and both groups.
 
1
  ---
 
 
 
2
  base_model:
3
  - Skywork/Matrix-Game
4
+ language:
5
+ - en
6
+ license: mit
7
+ library_name: diffusers
8
+ pipeline_tag: image-to-video
9
  ---
10
+
11
  <!-- markdownlint-disable first-line-h1 -->
12
  <!-- markdownlint-disable html -->
13
  <!-- markdownlint-disable no-duplicate-header -->
 
20
  <a href="https://arxiv.org/pdf/2506.18701">
21
  <img src="https://img.shields.io/badge/arXiv-Report-b31b1b?style=flat&logo=arxiv&logoColor=white" alt="arXiv">
22
  </a>
23
+ <a href="https://matrix-game-homepage.github.io/">
24
+ <img src="https://img.shields.io/badge/Project%20Page-grey?style=flat&logo=huggingface&color=FFA500" alt="Project Page">
25
+ </a>
26
 
27
 
28
  </div>
 
43
  ## πŸš€ Performance Comparison
44
  ### GameWorld Score Benchmark Comparison
45
 
46
+ | Model | Image Quality ↑ | Aesthetic Quality ↑ | Temporal Cons. ↑ | Motion Smooth. ↑ | Keyboard Acc. ↑ | Mouse Acc. ↑ | Object Cons. | Scenario Cons.|
47
+ |-----------|------------------|-------------|-------------------|-------------------|------------------|---------------|-------------|-------------|
48
+ | Oasis | 0.65 | 0.48 | 0.94 | **0.98** | 0.77 | 0.56 | 0.56 | 0.86 |
49
+ | MineWorld | 0.69 | 0.47 | 0.95 | **0.98** | 0.86 | 0.64 | 0.51 | 0.92 |
50
+ | **Ours** | **0.72** | **0.49** | **0.97** | **0.98** | **0.95** | **0.95** | **0.76** | **0.93** |
51
 
52
  **Metric Descriptions**:
53
 
54
  - **Image Quality** / **Aesthetic**: Visual fidelity and perceptual appeal of generated frames
55
  - **Temporal Consistency** / **Motion Smoothness**: Temporal coherence and smoothness between frames
56
  - **Keyboard Accuracy** / **Mouse Accuracy**: Accuracy in following user control signals
57
+ - **Object Consistency**: Geometric stability and consistency of objects over time
58
+ - **Scenario Consistency**: Scenario consistency over time
59
 
60
  Please check our [GameWorld](https://github.com/SkyworkAI/Matrix-Game/tree/main/GameWorldScore) benchmark for detailed implementation.
61
 
62
  ### Human Evaluation
63
 
64
+ ![Human Win Rate](assets/imgs/human_win_rate.png)
65
 
66
  > Double-blind human evaluation by two independent groups across four key dimensions: **Overall Quality**, **Controllability**, **Visual Quality**, and **Temporal Consistency**.
67
  > Scores represent the percentage of pairwise comparisons in which each method was preferred. Matrix-Game consistently outperforms prior models across all metrics and both groups.