#81

# dataset #81
path: /capstor/store/cscs/swissai/infra01/vision-datasets/tokenized/alignment/mllm_dpo
kind: manual | status: active
samples: n/a | size: 441.1 MB | files: 14

fields:
- prompt_id: text
- n_images: int
- image_tok: int
- seq_chosen_len: int
- seq_rejected_len: int

tree (truncated):
```
binidx/lengths_train.parquet
index_train.parquet
index_validation.parquet
manifest.json
raw/train-00000.parquet
raw/validation-00000.parquet
tokens/train-00000.i32
tokens/validation-00000.i32
train.bin
train.idx
validation.bin
validation.idx
views/train-00000.parquet
views/validation-00000.parquet
```

samples

1

mllm-dpo-lrv-2389

1

681

741

723

2

mllm-dpo-lrv-435

1

681

755

782

3

mllm-dpo-scigraph-1128

1

841

1022

987

4

mllm-dpo-scigraph-771

1

777

1019

901

5

mllm-dpo-lrv-232

1

753

915

930

6

mllm-dpo-lrv-2050

1

553

656

644

7

mllm-dpo-scigraph-2612

1

331

479

438

8

mllm-dpo-scigraph-158

1

278

418

421

9

mllm-dpo-lrv-389

1

681

758

733

10

mllm-dpo-lrv-1884

1

691

826

852