CoolFace
Datasetpublic

pufanyi/VBVR-Bench

VBVR-Bench Re-hosted copy of Video-Reason/VBVR-Bench-Data, converted to standard HuggingFace parquet format. Splits in_domain: 50 tasks x 5 samples = 250 entries (tasks overlap with the VBVR training set). out_of_domain: 50 tasks x 5 samples = 250 entries (held-out reasoning tasks). Schema field type notes task_name string e.g. G-13_grid_number_sequence_data-generator video_idx string zero-padded sample id (00000..00004) domain… See the full description on the dataset page: https://huggingface.co/datasets/pufanyi/VBVR-Bench.

sourceHugging Faceapache-2.0updated 5mo agoView on Hugging Face
0likes24downloads
Dataset Card

VBVR-Bench

Re-hosted copy of Video-Reason/VBVR-Bench-Data, converted to standard HuggingFace parquet format.

Splits

  • —`in_domain`: 50 tasks x 5 samples = 250 entries (tasks overlap with the VBVR training set).
  • —`out_of_domain`: 50 tasks x 5 samples = 250 entries (held-out reasoning tasks).

Schema

fieldtypenotes
task_namestringe.g. G-13_grid_number_sequence_data-generator
video_idxstringzero-padded sample id (00000..00004)
domainstringduplicates split name; convenient for filtering
promptstringtask description fed to the I2V model
first_frameImage (PNG)I2V condition frame
final_frameImage (PNG)expected final frame
ground_truth_videobinary (MP4)reference video — decode with decord / PyAV

Quick load

python
from datasets import load_dataset

ds = load_dataset("pufanyi/VBVR-Bench", split="in_domain")
sample = ds[0]
sample["first_frame"]          # PIL.Image
sample["prompt"]               # str
sample["ground_truth_video"]   # raw MP4 bytes

# Decode the video with decord
import decord, io
vr = decord.VideoReader(io.BytesIO(sample["ground_truth_video"]))

Links