Datasets:
Tasks:
Text Generation
Modalities:
Text
Formats:
json
Languages:
English
Size:
1K - 10K
Tags:
multimodel-small
Add dataset card
Browse files
README.md
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language:
|
| 3 |
+
- en
|
| 4 |
+
task_categories:
|
| 5 |
+
- text-generation
|
| 6 |
+
tags:
|
| 7 |
+
- multimodel-small
|
| 8 |
+
---
|
| 9 |
+
|
| 10 |
+
# MultiModel Small Dataset
|
| 11 |
+
|
| 12 |
+
JSONL datasets used for the MultiModel Small training experiments.
|
| 13 |
+
|
| 14 |
+
## Files
|
| 15 |
+
|
| 16 |
+
- `pretrain.jsonl`: text pretraining samples
|
| 17 |
+
- `val.jsonl`: validation samples
|
| 18 |
+
- `instruct_train.jsonl`: instruction and image-description text
|
| 19 |
+
- `align_train.jsonl`: image-text alignment samples
|
| 20 |
+
|
| 21 |
+
The image references in the instruction and alignment files are illustrative unless the corresponding image files are provided separately.
|