flickr8k
Datasets
All datasets matching “flickr8k”flickr8k
Dataset Card for "flickr8k"
More Information needed
flickr8k
Flickr8k Captions With Splits
This dataset repackages the local Flickr8k image-caption corpus into a Hugging Face imagefolder layout with per-split metadata.csv files.
Structure
train/, dev/, and test/ each contain image files and a metadata.csv.
Each row represents one image with five captions.
The image feature is inferred from file_name in each metadata.csv.
Columns
image: image file loaded by the Hugging Face Datasets library
file_name: image filename… See the full description on the dataset page: https://huggingface.co/datasets/intro/flickr8k.flickr8k
How created?
!kaggle datasets download -d adityajn105/flickr8k
!unzip -q /content/flickr8k.zip
import pandas as pd
from datasets import Dataset, Image
df = pd.read_csv("/content/captions.txt")
df["image"] = df["image"].apply(lambda x: f"/content/Images/{x}")
dataset = Dataset.from_pandas(df).cast_column("image", Image())
dataset.push_to_hub("ariG23498/flickr8k")
wds_flickr8kFlickr8k-ImagesSynCap-Flickr8k
SynCap-Flickr8k
Overview
The SynCap-Flickr8k is a unique collection designed to benchmark Vision Language Models (VLMs) in the image captioning task. This dataset was created using the Flickr8k dataset, which contains 8,000 images, each accompanied by five human-generated captions. By leveraging advanced models such as GPT-4o and LLaMA 3.2, we generated synthetic captions that enhance the understanding of how well VLMs can interpret and describe visual content.… See the full description on the dataset page: https://huggingface.co/datasets/kargwalaryan/SynCap-Flickr8k.