vogue-fashion-collection-15-nobg

Model description

This model classifies an image into a fashion collection. It is trained on the tonyassi/vogue-runway-top15-512px-nobg dataset and fine-tuned version of google/vit-base-patch16-224-in21k.

Because the model trained on a dataset with white background it is suggested to only give the model an image with a white background. Removing the background allows the model to focus on the clothes and disregard the background.

Dataset description

tonyassi/vogue-runway-top15-512px-nobg

• 15 fashion houses
• 1679 collections
• 87,547 images
• No background

How to use

from transformers import pipeline

# Initialize image classification pipeline
pipe = pipeline("image-classification", model="tonyassi/vogue-fashion-collection-15-nobg")

# Perform classification
result = pipe('image.png')

# Print results
print(result)

Training and evaluation data

It achieves the following results on the evaluation set:

• Loss: 0.5880
• Accuracy: 0.8403

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 10

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu121
• Datasets 2.16.1
• Tokenizers 0.15.0