🫁 Add fine-tuned ViT-Base lung cancer classifier (normal/malignant/benign)

README.md

@@ -0,0 +1,92 @@
+---
+license: apache-2.0
+tags:
+  - image-classification
+  - vision-transformer
+  - vit
+  - lung-cancer
+  - medical-imaging
+  - pytorch
+  - transformers
+base_model: google/vit-base-patch16-224
+pipeline_tag: image-classification
+---
+
+# 🫁 ViT Lung Cancer Classifier
+
+Fine-tuned **Vision Transformer (ViT-Base/16)** for lung cancer CT image classification
+into 3 classes: **normal**, **malignant**, and **benign**.
+
+## 📊 Model Details
+
+| Property | Value |
+|---|---|
+| Base Model | `google/vit-base-patch16-224` |
+| Task | Image Classification (3 classes) |
+| Input Size | 224 × 224 px |
+| Precision | fp16 |
+| Training | Full fine-tuning + early stopping |
+
+## 🏷️ Label Mapping
+
+| ID | Label | Description |
+|---|---|---|
+| 0 | `normal` | Normal lung tissue |
+| 1 | `malignant` | Malignant (cancerous) tissue |
+| 2 | `benign` | Benign (non-cancerous) tissue |
+
+## 🚀 Usage
+
+### Install
+
+```bash
+pip install transformers torch pillow
+
+### Python Inference
+Inference
+
+```python
+from transformers import ViTForImageClassification, ViTImageProcessor
+from PIL import Image
+import torch
+
+model_id = "TurkishCodeMan/vit-lung-cancer"
+
+processor = ViTImageProcessor.from_pretrained(model_id)
+model     = ViTForImageClassification.from_pretrained(model_id)
+
+device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+model.eval().to(device)
+
+def predict(image_path: str) -> dict:
+    img    = Image.open(image_path).convert("RGB")
+    inputs = processor(images=img, return_tensors="pt").to(device)
+
+    with torch.no_grad():
+        logits = model(**inputs).logits
+
+    pred_id = logits.argmax(-1).item()
+    probs   = torch.softmax(logits.float(), dim=-1)[0]
+
+    return {
+        "prediction": model.config.id2label[pred_id],
+        "probabilities": {
+            label: round(probs[i].item(), 4)
+            for i, label in model.config.id2label.items()
+        }
+    }
+
+result = predict("lung_scan.jpg")
+print(result)
+# {'prediction': 'malignant', 'probabilities': {'normal': 0.02, 'malignant': 0.91, 'benign': 0.07}}
+
+🛠️ Training Config
+
+Parameter: Value
+Optimizer: AdamW
+Learning Rate: 2e-5
+Batch Size: 16
+Max Epochs: 30
+Early Stopping Patience: 5
+Mixed Precision: fp16
+Best Metric: F1-Macro

config.json

@@ -0,0 +1,36 @@
+{
+  "architectures": [
+    "ViTForImageClassification"
+  ],
+  "attention_probs_dropout_prob": 0.0,
+  "dtype": "float32",
+  "encoder_stride": 16,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.0,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "normal",
+    "1": "malignant",
+    "2": "benign"
+  },
+  "image_size": 224,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "benign": 2,
+    "malignant": 1,
+    "normal": 0
+  },
+  "layer_norm_eps": 1e-12,
+  "model_type": "vit",
+  "num_attention_heads": 12,
+  "num_channels": 3,
+  "num_hidden_layers": 12,
+  "patch_size": 16,
+  "pooler_act": "tanh",
+  "pooler_output_size": 768,
+  "problem_type": "single_label_classification",
+  "qkv_bias": true,
+  "transformers_version": "5.2.0",
+  "use_cache": false
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a88dd7557323a7795db5729f781bbc440d92ebc487027c117bf10a5e17cb9189
+size 343227052

preprocessor_config.json

@@ -0,0 +1,23 @@
+{
+  "do_convert_rgb": null,
+  "do_normalize": true,
+  "do_rescale": true,
+  "do_resize": true,
+  "image_mean": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "image_processor_type": "ViTImageProcessor",
+  "image_std": [
+    0.5,
+    0.5,
+    0.5
+  ],
+  "resample": 2,
+  "rescale_factor": 0.00392156862745098,
+  "size": {
+    "height": 224,
+    "width": 224
+  }
+}