Datasets:

yurkes
/

patch_tasks_vllm

Tasks:

Image-Text-to-Text

Visual Question Answering

Modalities:

Formats:

Languages:

Size:

Libraries:

License:

Dataset card Data Studio Files Files and versions

yurkes commited on Oct 3, 2025

Commit

3dcdcc1

·

verified ·

1 Parent(s): 3f99af6

Update README.md

Files changed (1) hide show

README.md +1 -1

README.md CHANGED Viewed

@@ -34,7 +34,7 @@ pretty_name: QA Patches Task Dataset
 ### Dataset Description
-This dataset contains approximately 160,000 triplets of `question`, `answer`, and `image` designed for patch-based visual reasoning tasks.
 A standard question in this dataset is formatted as follows:

 ### Dataset Description
+This dataset contains approximately 305,000 triplets of `question`, `answer`, and `image` designed for patch-based visual reasoning tasks.
 A standard question in this dataset is formatted as follows: