Datasets:
read me added
Browse files
README.md
CHANGED
|
@@ -79,7 +79,7 @@ The tampering effects are applied to the middle of each video to ensure consiste
|
|
| 79 |
|
| 80 |
## Dataset Details
|
| 81 |
|
| 82 |
-
The
|
| 83 |
|
| 84 |
- **Content Diversity**: Spanning a variety of objects, activities, and settings.
|
| 85 |
- **Temporal Dynamics**: Videos with temporal dependencies for coherence testing.
|
|
@@ -87,7 +87,7 @@ The MVTamperBench dataset is built upon the **MVBench dataset**, a widely recogn
|
|
| 87 |
|
| 88 |
### Incorporated Datasets
|
| 89 |
|
| 90 |
-
The
|
| 91 |
|
| 92 |
| Dataset Name | Primary Scene Type and Unique Characteristics |
|
| 93 |
|----------------------|-------------------------------------------------------------------------|
|
|
@@ -129,9 +129,9 @@ By maintaining consistent tampering duration (1 second) and location (center of
|
|
| 129 |
|
| 130 |
## Download Dataset
|
| 131 |
|
| 132 |
-
You can access the
|
| 133 |
|
| 134 |
-
[Download
|
| 135 |
|
| 136 |
---
|
| 137 |
|
|
@@ -139,7 +139,7 @@ You can access the MVTamperBench dataset directly from the Hugging Face reposito
|
|
| 139 |
|
| 140 |
1. Clone the Hugging Face repository:
|
| 141 |
```bash
|
| 142 |
-
git clone [https://huggingface.co/datasets/
|
| 143 |
cd mvtamperbench
|
| 144 |
```
|
| 145 |
|
|
@@ -147,7 +147,7 @@ You can access the MVTamperBench dataset directly from the Hugging Face reposito
|
|
| 147 |
```python
|
| 148 |
from datasets import load_dataset
|
| 149 |
|
| 150 |
-
dataset = load_dataset("
|
| 151 |
```
|
| 152 |
|
| 153 |
3. Explore the dataset structure and metadata:
|
|
|
|
| 79 |
|
| 80 |
## Dataset Details
|
| 81 |
|
| 82 |
+
The MVTamperBenchEnd dataset is built upon the **MVBench dataset**, a widely recognized collection used in video-language evaluation. It features a broad spectrum of content to ensure robust model evaluation, including:
|
| 83 |
|
| 84 |
- **Content Diversity**: Spanning a variety of objects, activities, and settings.
|
| 85 |
- **Temporal Dynamics**: Videos with temporal dependencies for coherence testing.
|
|
|
|
| 87 |
|
| 88 |
### Incorporated Datasets
|
| 89 |
|
| 90 |
+
The MVTamperBenchEnd dataset integrates videos from several sources, each contributing unique characteristics:
|
| 91 |
|
| 92 |
| Dataset Name | Primary Scene Type and Unique Characteristics |
|
| 93 |
|----------------------|-------------------------------------------------------------------------|
|
|
|
|
| 129 |
|
| 130 |
## Download Dataset
|
| 131 |
|
| 132 |
+
You can access the MVTamperBenchEnd dataset directly from the Hugging Face repository:
|
| 133 |
|
| 134 |
+
[Download MVTamperBenchEnd Dataset](https://huggingface.co/datasets/Srikant86/MVTamperBenchEnd)
|
| 135 |
|
| 136 |
---
|
| 137 |
|
|
|
|
| 139 |
|
| 140 |
1. Clone the Hugging Face repository:
|
| 141 |
```bash
|
| 142 |
+
git clone [https://huggingface.co/datasets/mvtamperbenchend](https://huggingface.co/datasets/Srikant86/MVTamperBenchEnd)
|
| 143 |
cd mvtamperbench
|
| 144 |
```
|
| 145 |
|
|
|
|
| 147 |
```python
|
| 148 |
from datasets import load_dataset
|
| 149 |
|
| 150 |
+
dataset = load_dataset("mvtamperbenchend")
|
| 151 |
```
|
| 152 |
|
| 153 |
3. Explore the dataset structure and metadata:
|