🫁 MobileViT-Small — Lung Cancer Histopathology (LoRA Fine-tune)
This model is a LoRA fine-tuned version of apple/mobilevit-small
for lung cancer classification using histopathological images from
Lung Cancer Histopathological Images Dataset.
🧩 Configuration
| Attribute | Value |
|---|---|
| Base Model | apple/mobilevit-small |
| Fine-tuning Method | LoRA (Low-Rank Adaptation) |
| Dataset | Lung Cancer Histopathological Images |
| Classes | Adenocarcinoma, Benign, Squamous Cell Carcinoma |
| Epochs | 10 |
| Batch Size | 32 |
| Learning Rate | 0.0005 |
| Optimizer | AdamW |
| LoRA Config | r=4, α=32, dropout=0.1, target_modules=[query, value] |
| Trainable Params | 30.3K / 4.97M (0.61%) |
🔬 Example Image
📊 Evaluation Results
| Metric | Avg | Adenocarcinoma | Benign | Squamous Cell Carcinoma |
|---|---|---|---|---|
| Accuracy | 0.9974 | 0.9961 | 1.0000 | 0.9961 |
| Precision | 0.9963 | 0.9889 | 1.0000 | 1.0000 |
| Recall | 0.9960 | 1.0000 | 1.0000 | 0.9880 |
| F1 Score | 0.9961 | 0.9944 | 1.0000 | 0.9939 |
| AUC | 1.0000 | 0.9999 | 1.0000 | 0.9999 |
Test loss: 0.00425 Inference time: 0.678 s