Datasets:
image_str stringlengths 169k 242k | input stringlengths 17 191 | outputs stringclasses 27
values | meta dict |
|---|---|---|---|
iVBORw0KGgoAAAANSUhEUgAAAeAAAAFACAYAAABkyK97AAEAAElEQVR4nIT96ZZkSXImiH2i1yOydhTQWBro5pzh4dtx3pcPMDwzJGfYAKpQS2aG21X+UPkWuRY4tKpIdze7pousn4iKqtb/8r/83/d9b1QBewOlX84Pvq7rwn2/AACFwsb5cO+NqoX7vlEFVC29D5z2zjPnmwCwqnDvHe8j2ipgA6jCvm+NY1WdtrDPuPZGreox7m7b34+h4+PjA6/XC9gb6zpj3fvM6fX61LiqB7P3JjHOnFDqoR/VGM74ctyHBmfsJZq8Xp+oWqgC... | Do the thing that is behind the brown metallic ball and the small metal thing that is behind the tiny yellow metallic cube have the same color? | yes | {
"image_index": 39065,
"question_index": 390645
} |
"iVBORw0KGgoAAAANSUhEUgAAAeAAAAFACAYAAABkyK97AAEAAElEQVR4nHz9ybZtS3IdiE3zte99FQIBBCoCDBIiKSLJVCpTY6i(...TRUNCATED) | "Are there the same number of big shiny cubes on the left side of the tiny shiny cube and big sphere(...TRUNCATED) | yes | {
"image_index": 1882,
"question_index": 18828
} |
"iVBORw0KGgoAAAANSUhEUgAAAeAAAAFACAYAAABkyK97AAEAAElEQVR4nIT9abN1SXYehj1rn1tVXT0CIEZKtkSHw2Ep7A92hP3(...TRUNCATED) | How many other metal blocks are the same color as the block? | 0 | {
"image_index": 5152,
"question_index": 51529
} |
"iVBORw0KGgoAAAANSUhEUgAAAeAAAAFACAYAAABkyK97AAEAAElEQVR4nIT9S5NtS3IeiH0ea+3Mc+69datQqAIKD6JBQo2m1Ma(...TRUNCATED) | Are there any gray cubes in front of the brown metal cube? | yes | {
"image_index": 37168,
"question_index": 371675
} |
"iVBORw0KGgoAAAANSUhEUgAAAeAAAAFACAYAAABkyK97AAEAAElEQVR4nIT9a5dkSZIciIna9cysV88MCIDA7p4lD3d/Hvcf85D(...TRUNCATED) | What is the material of the large ball behind the cyan matte cylinder? | rubber | {
"image_index": 31322,
"question_index": 313211
} |
"iVBORw0KGgoAAAANSUhEUgAAAeAAAAFACAYAAABkyK97AAEAAElEQVR4nHz9y5JtS7IlCA21udz33iee91beJB+XSsikGtWBKhq(...TRUNCATED) | How many other objects are the same color as the big matte cylinder? | 1 | {
"image_index": 13767,
"question_index": 137666
} |
"iVBORw0KGgoAAAANSUhEUgAAAeAAAAFACAYAAABkyK97AAEAAElEQVR4nHT9244tSZIliC1R2+7nEhkZea9k9XS/zBDgECAwwGA(...TRUNCATED) | What size is the red block? | large | {
"image_index": 68755,
"question_index": 687542
} |
"iVBORw0KGgoAAAANSUhEUgAAAeAAAAFACAYAAABkyK97AAEAAElEQVR4nIT92bJsSZIdiC21fW5E5AwkIASk2W/kC/lE/hZ+t0V(...TRUNCATED) | "There is a matte object behind the rubber thing that is on the left side of the large red object; w(...TRUNCATED) | large | {
"image_index": 5669,
"question_index": 56694
} |
"iVBORw0KGgoAAAANSUhEUgAAAeAAAAFACAYAAABkyK97AAEAAElEQVR4nIT9ybJlS5IdiC21fdz9NdFkZmQigUwgUYICKKiiCGe(...TRUNCATED) | How many big metal objects are in front of the big purple object? | 0 | {
"image_index": 14740,
"question_index": 147400
} |
"iVBORw0KGgoAAAANSUhEUgAAAeAAAAFACAYAAABkyK97AAEAAElEQVR4nHz9aZcsy5EkBoqaR963AA+FLlSxeZqHQ575ef2b55D(...TRUNCATED) | What number of brown objects are either big blocks or large rubber objects? | 0 | {
"image_index": 5692,
"question_index": 56921
} |
M3IT: five selected training subsets
Unmodified copies of five training JSONL files from
MMInstruction/M3IT.
Source revision: 2256415bd85b14945ac717170274565adb2a2b2d.
Included subsets: CLEVR, Visual-MRC, VIQuAE, OCR-VQA, and ST-VQA.
Original relative paths are preserved. VIQuAE uses train_v2.jsonl;
CLEVR uses the ordinary train.jsonl.
| File | Bytes |
|---|---|
data/reasoning/clevr/train.jsonl |
5,967,894,063 |
data/reasoning/visual_mrc/train.jsonl |
4,240,512,676 |
data/vqa/viquae/train_v2.jsonl |
115,592,369 |
data/vqa/ocr-vqa/train.jsonl |
559,430,958 |
data/vqa/st-vqa/train.jsonl |
1,478,233,045 |
Total data size: 12,361,663,111 bytes (12.36 GB).
See source_manifest.json for the source revision, sizes, and SHA256 hashes.
Only these five source files are included; validation/test splits, alternative
training versions, the original dataset loader, and instructions.json are omitted.
This repository preserves raw JSONL files; it does not reproduce the original
M3IT loader's instruction assembly.
Download
hf download axiong/M3IT-selected-5-train --repo-type dataset --local-dir ./M3IT-selected-5-train
Attribution and terms
Source: Lei Li et al., M³IT: A Large-Scale Dataset towards Multi-Modal Multilingual Instruction Tuning (2023), arXiv:2306.04387. This is an unofficial subset mirror; the data files have not been modified. According to the source dataset card, original dataset content retains its original licenses, and M3IT's annotated instruction data is licensed under CC BY 4.0. This mirror grants no additional rights.
- Downloads last month
- 40