skin-lesion-models

Skin Lesion Classification Models

This repository contains trained PyTorch checkpoints for skin lesion classification on the PAD-UFES-20 dataset.

📊 Models

🚀 Usage

from huggingface_hub import hf_hub_download
import torch

# Download a specific model
ckpt_path = hf_hub_download(
    repo_id="muhwira27/skin-lesion-models",
    filename="shufflenet_v2_x1_0_best.ckpt"
)

# Load checkpoint
checkpoint = torch.load(ckpt_path, map_location="cpu")

# Checkpoint contents
# - backbone: str (model architecture name)
# - state_dict: model weights
# - label2id: dict mapping class names to indices
# - img_size: int (input image size, typically 224)

📦 Available Files

• shufflenet_v2_x1_0_best.ckpt - Best model (smallest & best performance)
• densenet121_best.ckpt
• vit_small_patch16_224_best.ckpt
• seresnet50_best.ckpt
• tf_efficientnetv2_s_best.ckpt
• resnet50_best.ckpt
• regnety_032_best.ckpt
• mobilenet_v3_large_best.ckpt
• convnext_tiny_best.ckpt
• efficientnet_b0_best.ckpt

🏥 Classes

• ACK: Actinic Keratosis
• BCC: Basal Cell Carcinoma
• MEL: Melanoma
• NEV: Nevus (Mole)
• SCC: Squamous Cell Carcinoma
• SEK: Seborrheic Keratosis

📄 License

MIT License

👤 Author

Muh. Wira Ramdhani Fadhil