tejasssuthrave/telidermai

vit_base_patch16_224-finetuned-SkinDisease

This model is a fine-tuned version of google/vit-base-patch16-224 on a custom skin disease dataset (image_folder format).
It achieves the following results on the evaluation set:

• Loss: 0.1992
• Accuracy: 0.9343

---

🧠 Model description

This model is a Vision Transformer (ViT) fine-tuned for skin disease classification.
It processes input images of size 224x224 pixels and outputs the most likely class.

---

🔗 Intended uses & limitations

• ✅ For clinical support, not for standalone medical diagnosis.
• ✅ Designed for educational, research, and proof-of-concept use.

---

🧪 Training and evaluation data

• Dataset used: Custom image dataset with labeled skin diseases
• Preprocessing: Resized to 224×224, normalized

---

⚙️ Training procedure

Hyperparameters:

• Learning Rate: 5e-05
• Epochs: 10
• Batch Size: 32
• Optimizer: Adam
• Scheduler: Linear with warmup
• Seed: 42

Training results:

Epoch	Val Loss	Accuracy
1	0.8248	0.7647
2	0.4236	0.8748
3	0.3154	0.9021
4	0.2695	0.9106
5	0.2381	0.9198
6	0.2407	0.9218
7	0.2160	0.9278
8	0.2121	0.9283
9	0.2044	0.9303
10	0.1992	0.9343

---

🧰 Framework versions

• transformers: 4.33.2
• pytorch: 2.0.0
• datasets: 2.1.0
• tokenizers: 0.13.3