Sapek007/banana-ripeness

Banana Ripeness Classification

Image dataset for classifying banana ripeness into four classes:

• unripe
• ripe
• overripe
• rotten

Sourced from Roboflow (Banana Ripeness Classification, v6).

Structure

The ZIP archive contains a standard train/valid/test split with one subdirectory per class:

banana_ripeness_dataset/
├── train/
│   ├── unripe/
│   ├── ripe/
│   ├── overripe/
│   └── rotten/
├── valid/   (same 4 subdirs)
└── test/    (same 4 subdirs)

Usage

from huggingface_hub import hf_hub_download
import zipfile, os

zip_path = hf_hub_download(
    repo_id="Sapek007/banana-ripeness",
    filename="banana_ripeness_dataset.zip",
    repo_type="dataset",
)

extract_dir = "banana_data"
if not os.path.exists(extract_dir):
    with zipfile.ZipFile(zip_path) as z:
        z.extractall(extract_dir)

# Use with PyTorch / Keras ImageFolder loaders
import torchvision
train_ds = torchvision.datasets.ImageFolder(f"{extract_dir}/train")

Citation

Compiled for an ML course final project at Copenhagen Business School (2026). Original dataset by Roboflow under CC-BY-4.0.