Update README.md
Browse files
README.md
CHANGED
|
@@ -18,6 +18,12 @@ Plus this could be a good benchmark for Google/OpenAI/etc to try and beat it.
|
|
| 18 |
|
| 19 |
[*Good luck!*](https://cdn-uploads.huggingface.co/production/uploads/633e85093a17ab61de8d9073/sDNCW79VG7kZEseJOF2yC.png)
|
| 20 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 21 |
## License
|
| 22 |
|
| 23 |
N/A
|
|
|
|
| 18 |
|
| 19 |
[*Good luck!*](https://cdn-uploads.huggingface.co/production/uploads/633e85093a17ab61de8d9073/sDNCW79VG7kZEseJOF2yC.png)
|
| 20 |
|
| 21 |
+
## Rough Idea of using this on benchmarking
|
| 22 |
+
|
| 23 |
+
1. Do not use the dataset for training
|
| 24 |
+
2. The model must be able to accept visual and audio input
|
| 25 |
+
3. When asking to describe the video, it should not produce any "made up" or "wrong / incorrect" facts.
|
| 26 |
+
|
| 27 |
## License
|
| 28 |
|
| 29 |
N/A
|