Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,49 @@
|
|
1 |
-
---
|
2 |
-
license: apache-2.0
|
3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: apache-2.0
|
3 |
+
task_categories:
|
4 |
+
- text-generation
|
5 |
+
language:
|
6 |
+
- en
|
7 |
+
- zh
|
8 |
+
size_categories:
|
9 |
+
- 1M<n<10M
|
10 |
+
---
|
11 |
+
# Dataset Card for Oryx-SFT-Data
|
12 |
+
|
13 |
+
This dataset is used for the training of the Oryx model. We only allow the use of this dataset for academic research and education purpose.
|
14 |
+
|
15 |
+
For more details, please refer to our [Github Repo](https://github.com/liuzuyan/oryx)
|
16 |
+
|
17 |
+
## Dataset Details
|
18 |
+
|
19 |
+
A mixture of 1.2M image/video data.
|
20 |
+
|
21 |
+
For more statistics of the dataset, please refer to our paper (coming soon)
|
22 |
+
|
23 |
+
|
24 |
+
### Source Data
|
25 |
+
|
26 |
+
We preprocess all the data into multiple files.
|
27 |
+
|
28 |
+
You can use our preprocessed data according to our [Github Repo](https://github.com/liuzuyan/oryx)
|
29 |
+
|
30 |
+
Or you can download raw data alternatively.
|
31 |
+
|
32 |
+
The links are as below:
|
33 |
+
|
34 |
+
VideoChatGPT-Plus: https://huggingface.co/datasets/MBZUAI/VCG-plus_112K
|
35 |
+
|
36 |
+
LLaVA-Hound: https://huggingface.co/ShareGPTVideo/LLaVA-Hound-SFT
|
37 |
+
|
38 |
+
ShareGPT4Video: https://huggingface.co/datasets/ShareGPT4Video/ShareGPT4Video
|
39 |
+
|
40 |
+
Cinepile: https://huggingface.co/datasets/tomg-group-umd/cinepile
|
41 |
+
|
42 |
+
NExT-QA: https://github.com/doc-doc/NExT-QA
|
43 |
+
|
44 |
+
Perception Test: https://github.com/google-deepmind/perception_test
|
45 |
+
|
46 |
+
ScanQA: https://github.com/ATR-DBI/ScanQA
|
47 |
+
|
48 |
+
|
49 |
+
## Citation
|