Update README.md
Browse files
README.md
CHANGED
|
@@ -50,8 +50,13 @@ The primary intended users of the model are researchers and hobbyists in compute
|
|
| 50 |
|
| 51 |
## Training dataset
|
| 52 |
595K filtered image-text pairs from CC3M.
|
|
|
|
| 53 |
150K GPT-generated multimodal instruction-following chat data.
|
|
|
|
| 54 |
83K VQA v2 instruction-following VQA data.
|
|
|
|
| 55 |
16K A-OKVQA instruction-following CoT-VQA data.
|
|
|
|
| 56 |
23K FLICKR instruction-following spotting captioning data.
|
|
|
|
| 57 |
10K LLaVA-based human preference data
|
|
|
|
| 50 |
|
| 51 |
## Training dataset
|
| 52 |
595K filtered image-text pairs from CC3M.
|
| 53 |
+
|
| 54 |
150K GPT-generated multimodal instruction-following chat data.
|
| 55 |
+
|
| 56 |
83K VQA v2 instruction-following VQA data.
|
| 57 |
+
|
| 58 |
16K A-OKVQA instruction-following CoT-VQA data.
|
| 59 |
+
|
| 60 |
23K FLICKR instruction-following spotting captioning data.
|
| 61 |
+
|
| 62 |
10K LLaVA-based human preference data
|