You can deploy this model and start earning money today!

raphael0202/
amenity-website-images

Name: amenity-website-images
Availability: InStock

ultralytics

onnx

image-classification

raphael0202/amenity-website-images

agpl-3.0

Views

No views yet

Amenity website image classifier

Wandb tracking run

This image classification model can be used to classify images crawled from amenity websites into a fixed set of categories:

amenity_inside (ID: 0): photos of the inside of the amenity, possibly with persons displayed
amenity_outside (ID: 1): photos of the outside of the amenity: can be possibly be a photo of the facade of the amenity.
food_and_drink (ID: 2): photos of the food or drinks that are served in the amenity (mostly for restaurants/cafe/hotels)
other (ID: 3): all other images, including icons, logos,...

The model was fine-tuned using the Ultralytics YOLO library.

Model Details

Model Description

Developed by: Raphaël Bournhonesque
Model type: image classification
License: agpl-3.0
Finetuned from model [optional]: yolov8n-cls.pt

Training Details

Training Data

The model was fine-tuned using the following dataset: raphael0202/amenity-website-images (revision: main). As the dataset contains mostly images crawled from websites of restaurants and coffee shop, the model may exhibit lower accuracy on other types of amenities.

Training Procedure

Dependency versions:

ultralytics: 8.4.14
pytorch: 2.9.0+cu128

Training Hyperparameters

Epochs: 100
Batch size: 16
Image size: 640

Evaluation

The following evaluation metrics were obtained after training the model:

metrics/accuracy_top1: 0.90183025598526
metrics/accuracy_top5: 1.0
fitness: 0.95091512799263

Evaluation on exported models

The model was also evaluated after exporting to ONNX format. The following metrics were obtained:

ONNX export

metrics/accuracy_top1: 0.90183025598526
metrics/accuracy_top5: 1.0
fitness: 0.95091512799263

Files

Most files stored on the repo are standard files created during training with the Ultralytics YOLO library.

What was added:

an ONNX export of the trained model (best model), stored in weights/model.onnx.
a Parquet file containing predictions on the full dataset, stored in predictions.parquet.
metrics JSON files for each exported model format, stored in metrics_*.json:
- metrics.json: metrics for the original PyTorch model
- metrics_onnx.json: metrics for the ONNX exported model