Update README.md
Browse files
README.md
CHANGED
|
@@ -30,11 +30,11 @@ BakLLaVA 2 is cooking with a significantly larger (commercially viable) dataset
|
|
| 30 |
|
| 31 |
# Training dataset
|
| 32 |
|
| 33 |
-
-558K filtered image-text pairs from LAION/CC/SBU, captioned by BLIP.
|
| 34 |
-
-158K GPT-generated multimodal instruction-following data.
|
| 35 |
-
-450K academic-task-oriented VQA data mixture.
|
| 36 |
-
-40K ShareGPT data.
|
| 37 |
-
-Additional private data (permissive)
|
| 38 |
|
| 39 |
|
| 40 |
|
|
|
|
| 30 |
|
| 31 |
# Training dataset
|
| 32 |
|
| 33 |
+
- 558K filtered image-text pairs from LAION/CC/SBU, captioned by BLIP.
|
| 34 |
+
- 158K GPT-generated multimodal instruction-following data.
|
| 35 |
+
- 450K academic-task-oriented VQA data mixture.
|
| 36 |
+
- 40K ShareGPT data.
|
| 37 |
+
- Additional private data (permissive)
|
| 38 |
|
| 39 |
|
| 40 |
|