Johnyquest7
/

TN5000_model

@@ -3,13 +3,21 @@ library_name: transformers
 license: apache-2.0
 base_model: microsoft/swinv2-base-patch4-window8-256
 tags:
 - generated_from_trainer
 - ml-intern
 datasets:
-- generator
 metrics:
 - accuracy
 - f1
 model-index:
 - name: TN5000_model
   results:
@@ -17,124 +25,346 @@ model-index:
       name: Image Classification
       type: image-classification
     dataset:
-      name: generator
-      type: generator
-      config: default
-      split: train
-      args: default
     metrics:
     - name: Accuracy
       type: accuracy
       value: 0.872
     - name: F1
       type: f1
-      value: 0.9080459770114943
 ---
-<!-- This model card has been generated automatically according to the information the Trainer had access to. You
-should probably proofread and complete it, then remove this comment. -->
-# TN5000_model
-This model is a fine-tuned version of [microsoft/swinv2-base-patch4-window8-256](https://huggingface.co/microsoft/swinv2-base-patch4-window8-256) on the generator dataset.
-It achieves the following results on the evaluation set:
-- Loss: 0.1293
-- Accuracy: 0.872
-- F1: 0.9080
-- Sensitivity: 0.8745
-- Specificity: 0.8654
-- Ppv: 0.9442
-- Npv: 0.7258
-- Auc Roc: 0.9367
-- Tp: 474
-- Tn: 180
-- Fp: 28
-- Fn: 68
-## Model description
-More information needed
-## Intended uses & limitations
-More information needed
-## Training and evaluation data
-More information needed
-## Training procedure
-### Training hyperparameters
-The following hyperparameters were used during training:
-- learning_rate: 2e-05
-- train_batch_size: 16
-- eval_batch_size: 16
-- seed: 42
-- gradient_accumulation_steps: 2
-- total_train_batch_size: 32
-- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
-- lr_scheduler_type: linear
-- lr_scheduler_warmup_steps: 0.1
-- num_epochs: 30
-### Training results
-| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1     | Sensitivity | Specificity | Ppv    | Npv    | Auc Roc | Tp  | Tn  | Fp | Fn  |
-|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|:-----------:|:-----------:|:------:|:------:|:-------:|:---:|:---:|:--:|:---:|
-| 0.3555        | 1.0   | 88   | 0.2034          | 0.6786   | 0.7604 | 0.714       | 0.59        | 0.8132 | 0.4521 | 0.7130  | 357 | 118 | 82 | 143 |
-| 0.2755        | 2.0   | 176  | 0.1604          | 0.7314   | 0.7920 | 0.716       | 0.77        | 0.8861 | 0.5203 | 0.8209  | 358 | 154 | 46 | 142 |
-| 0.2580        | 3.0   | 264  | 0.1648          | 0.8143   | 0.8684 | 0.858       | 0.705       | 0.8791 | 0.6651 | 0.8609  | 429 | 141 | 59 | 71  |
-| 0.2399        | 4.0   | 352  | 0.1343          | 0.7686   | 0.8208 | 0.742       | 0.835       | 0.9183 | 0.5642 | 0.8774  | 371 | 167 | 33 | 129 |
-| 0.2292        | 5.0   | 440  | 0.1547          | 0.8757   | 0.9153 | 0.94        | 0.715       | 0.8918 | 0.8266 | 0.9116  | 470 | 143 | 57 | 30  |
-| 0.2285        | 6.0   | 528  | 0.1153          | 0.8186   | 0.8656 | 0.818       | 0.82        | 0.9191 | 0.6431 | 0.9161  | 409 | 164 | 36 | 91  |
-| 0.2417        | 7.0   | 616  | 0.1111          | 0.8171   | 0.8621 | 0.8         | 0.86        | 0.9346 | 0.6324 | 0.9171  | 400 | 172 | 28 | 100 |
-| 0.1748        | 8.0   | 704  | 0.1145          | 0.8514   | 0.8919 | 0.858       | 0.835       | 0.9286 | 0.7017 | 0.9271  | 429 | 167 | 33 | 71  |
-| 0.1901        | 9.0   | 792  | 0.1489          | 0.8857   | 0.9206 | 0.928       | 0.78        | 0.9134 | 0.8125 | 0.9208  | 464 | 156 | 44 | 36  |
-| 0.1872        | 10.0  | 880  | 0.1162          | 0.8514   | 0.8919 | 0.858       | 0.835       | 0.9286 | 0.7017 | 0.9218  | 429 | 167 | 33 | 71  |
-| 0.1439        | 11.0  | 968  | 0.1108          | 0.8014   | 0.8478 | 0.774       | 0.87        | 0.9370 | 0.6063 | 0.9259  | 387 | 174 | 26 | 113 |
-| 0.1868        | 12.0  | 1056 | 0.1185          | 0.8714   | 0.9076 | 0.884       | 0.84        | 0.9325 | 0.7434 | 0.9320  | 442 | 168 | 32 | 58  |
-| 0.2002        | 13.0  | 1144 | 0.1376          | 0.8857   | 0.9205 | 0.926       | 0.785       | 0.9150 | 0.8093 | 0.9317  | 463 | 157 | 43 | 37  |
-| 0.2023        | 14.0  | 1232 | 0.1339          | 0.8857   | 0.9195 | 0.914       | 0.815       | 0.9251 | 0.7913 | 0.9350  | 457 | 163 | 37 | 43  |
-| 0.1582        | 15.0  | 1320 | 0.1346          | 0.8929   | 0.9252 | 0.928       | 0.805       | 0.9225 | 0.8173 | 0.9338  | 464 | 161 | 39 | 36  |
-| 0.1488        | 16.0  | 1408 | 0.1319          | 0.8957   | 0.9266 | 0.922       | 0.83        | 0.9313 | 0.8098 | 0.9366  | 461 | 166 | 34 | 39  |
-| 0.1249        | 17.0  | 1496 | 0.1280          | 0.87     | 0.9061 | 0.878       | 0.85        | 0.9360 | 0.7359 | 0.9370  | 439 | 170 | 30 | 61  |
-| 0.1553        | 18.0  | 1584 | 0.1121          | 0.8571   | 0.8943 | 0.846       | 0.885       | 0.9484 | 0.6969 | 0.9390  | 423 | 177 | 23 | 77  |
-| 0.1083        | 19.0  | 1672 | 0.1675          | 0.9029   | 0.9331 | 0.948       | 0.79        | 0.9186 | 0.8587 | 0.9381  | 474 | 158 | 42 | 26  |
-| 0.1379        | 20.0  | 1760 | 0.1535          | 0.9      | 0.9308 | 0.942       | 0.795       | 0.9199 | 0.8457 | 0.9340  | 471 | 159 | 41 | 29  |
-| 0.1458        | 21.0  | 1848 | 0.1915          | 0.9043   | 0.9344 | 0.954       | 0.78        | 0.9155 | 0.8715 | 0.9345  | 477 | 156 | 44 | 23  |
-| 0.1322        | 22.0  | 1936 | 0.1244          | 0.8786   | 0.9126 | 0.888       | 0.855       | 0.9387 | 0.7533 | 0.9399  | 444 | 171 | 29 | 56  |
-| 0.1438        | 23.0  | 2024 | 0.1519          | 0.8757   | 0.9122 | 0.904       | 0.805       | 0.9206 | 0.7703 | 0.9356  | 452 | 161 | 39 | 48  |
-| 0.1265        | 24.0  | 2112 | 0.1421          | 0.8871   | 0.9204 | 0.914       | 0.82        | 0.9270 | 0.7923 | 0.9378  | 457 | 164 | 36 | 43  |
-| 0.1118        | 25.0  | 2200 | 0.2165          | 0.9057   | 0.9358 | 0.962       | 0.765       | 0.9110 | 0.8895 | 0.9380  | 481 | 153 | 47 | 19  |
-| 0.0971        | 26.0  | 2288 | 0.1557          | 0.8857   | 0.9194 | 0.912       | 0.82        | 0.9268 | 0.7885 | 0.9342  | 456 | 164 | 36 | 44  |
-| 0.1235        | 27.0  | 2376 | 0.1394          | 0.88     | 0.9143 | 0.896       | 0.84        | 0.9333 | 0.7636 | 0.9376  | 448 | 168 | 32 | 52  |
-### Framework versions
-- Transformers 5.8.0
-- Pytorch 2.11.0+cu130
-- Datasets 4.8.5
-- Tokenizers 0.22.2
-<!-- ml-intern-provenance -->
-## Generated by ML Intern
-This model repository was generated by [ML Intern](https://github.com/huggingface/ml-intern), an agent for machine learning research and development on the Hugging Face Hub.
-- Try ML Intern: https://smolagents-ml-intern.hf.space
-- Source code: https://github.com/huggingface/ml-intern
-## Usage
 ```python
-from transformers import AutoModelForCausalLM, AutoTokenizer
-model_id = 'Johnyquest7/TN5000_model'
-tokenizer = AutoTokenizer.from_pretrained(model_id)
-model = AutoModelForCausalLM.from_pretrained(model_id)
 ```
-For non-causal architectures, replace `AutoModelForCausalLM` with the appropriate `AutoModel` class.

 license: apache-2.0
 base_model: microsoft/swinv2-base-patch4-window8-256
 tags:
+- image-classification
+- medical-imaging
+- thyroid
+- ultrasound
+- swinv2
 - generated_from_trainer
 - ml-intern
 datasets:
+- sadib2026/roi-dataset-tn5000
 metrics:
 - accuracy
 - f1
+- auc_roc
+- sensitivity
+- specificity
 model-index:
 - name: TN5000_model
   results:
       name: Image Classification
       type: image-classification
     dataset:
+      name: TN5000 ROI Dataset
+      type: image-classification
     metrics:
     - name: Accuracy
       type: accuracy
       value: 0.872
     - name: F1
       type: f1
+      value: 0.908
+    - name: AUC-ROC
+      type: auc_roc
+      value: 0.937
+    - name: Sensitivity
+      type: sensitivity
+      value: 0.875
+    - name: Specificity
+      type: specificity
+      value: 0.865
+---
+# TN5000 Thyroid Ultrasound Classifier
+**Fine-tuned SwinV2-Base for Benign vs Malignant Thyroid Nodule Classification**
+[![Model](https://img.shields.io/badge/Model-HuggingFace-yellow)](https://huggingface.co/Johnyquest7/TN5000_model)
+[![Demo](https://img.shields.io/badge/Demo-Gradio-green)](https://huggingface.co/spaces/Johnyquest7/tn5000-thyroid-demo)
+[![Dataset](https://img.shields.io/badge/Dataset-Kaggle-blue)](https://www.kaggle.com/datasets/sadib2026/roi-dataset-tn5000)
+---
+## 📋 Table of Contents
+1. [Overview](#overview)
+2. [Model Architecture](#model-architecture)
+3. [Dataset](#dataset)
+4. [Training Methodology](#training-methodology)
+5. [Results](#results)
+6. [External Validation](#external-validation)
+7. [How to Use](#how-to-use)
+8. [Limitations & Disclaimers](#limitations--disclaimers)
+9. [Citation](#citation)
+---
+## Overview
+This model classifies thyroid ultrasound images as **benign** or **malignant**, designed to assist in the early detection of thyroid cancer. It was fine-tuned from Microsoft's SwinV2-Base vision transformer on the TN5000 ROI dataset from Kaggle.
+**Key Design Decisions:**
+- **Optimized for sensitivity** (87.5%) to minimize missed malignancies — critical in cancer screening
+- **AUC-ROC of 0.94** indicates excellent discriminative ability
+- **Focal loss with class weighting** handles the benign/malignant class imbalance
+- **Early stopping** prevents overfitting on the small medical dataset
+---
+## Model Architecture
+| Property | Value |
+|----------|-------|
+| **Base Model** | [microsoft/swinv2-base-patch4-window8-256](https://huggingface.co/microsoft/swinv2-base-patch4-window8-256) |
+| **Architecture** | Swin Transformer V2 |
+| **Parameters** | 86.9M |
+| **Input Size** | 256 × 256 |
+| **Patch Size** | 4 × 4 |
+| **Window Size** | 8 × 8 |
+| **Number of Classes** | 2 (benign, malignant) |
+| **License** | Apache 2.0 |
+**Why SwinV2?**
+Swin Transformers use hierarchical feature maps and shifted window attention, making them particularly effective for medical imaging where local texture patterns (echogenicity, microcalcifications, irregular margins) are diagnostically important. SwinV2 improves training stability with a cosine attention mechanism and larger model capacity.
+---
+## Dataset
+### Primary Training Dataset: TN5000 ROI
+- **Source:** [Kaggle - ROI Dataset TN5000](https://www.kaggle.com/datasets/sadib2026/roi-dataset-tn5000)
+- **Type:** Thyroid ultrasound Region-of-Interest (ROI) patches
+- **Total Images:** 4,250
+| Split | Images | Benign | Malignant |
+|-------|--------|--------|-----------|
+| Train (80%) | 2,800 | ~1,600 | ~1,200 |
+| Validation (20%) | 700 | ~400 | ~300 |
+| Test (held-out) | 750 | ~400 | ~350 |
+**Class Distribution:** The dataset is moderately imbalanced with more benign cases. We used **balanced class weights** (benign: 1.75, malignant: 0.70) and **focal loss** (γ=2.0) to prioritize malignant case detection.
+### External Validation Dataset
+- **Source:** [Johnyquest7/thyroid-cancer-classification-ultrasound-dataset](https://huggingface.co/datasets/Johnyquest7/thyroid-cancer-classification-ultrasound-dataset)
+- **Images:** 3,115 total (train + test splits)
+- **Purpose:** Independent validation on unseen data from a different source
+---
+## Training Methodology
+### Data Preprocessing
+| Transform | Training | Validation/Test |
+|-----------|----------|-----------------|
+| Resize | RandomResizedCrop(256) | Resize(256) + CenterCrop(256) |
+| Horizontal Flip | 50% probability | No |
+| Rotation | ±10° | No |
+| Color Jitter | brightness=0.2, contrast=0.2 | No |
+| Normalization | ImageNet mean/std | ImageNet mean/std |
+### Training Configuration
+```python
+learning_rate: 2e-5
+batch_size: 16 (per device)
+gradient_accumulation_steps: 2
+effective_batch_size: 32
+epochs: 30 (early stopping patience: 5)
+warmup_ratio: 0.1
+optimizer: AdamW (β1=0.9, β2=0.999)
+scheduler: Linear with warmup
+mixed_precision: bf16
+seed: 42
+```
+### Loss Function: Focal Loss
+Standard cross-entropy treats all misclassifications equally. In thyroid screening, **missing a malignant case (false negative) is far more costly** than a false alarm. We used focal loss:
+```
+FL(pt) = −(1 − pt)^γ · log(pt)
+```
+With γ=2.0, the model focuses learning on hard-to-classify malignant cases. Class weights further upweight the minority malignant class.
+### Model Selection Criterion
+The best model was selected by **validation AUC-ROC** (not accuracy), ensuring optimal discrimination between benign and malignant cases across all thresholds.
+---
+## Results
+### Validation Set (700 images)
+| Metric | Value |
+|--------|-------|
+| **Accuracy** | 87.9% |
+| **F1-Score** | 91.3% |
+| **Sensitivity (Recall)** | 88.8% |
+| **Specificity** | 85.5% |
+| **PPV (Precision)** | 93.9% |
+| **NPV** | 75.3% |
+| **AUC-ROC** | **0.940** |
+**Confusion Matrix:**
+```
+              Predicted
+           Benign  Malignant
+Actual Benign    171       29
+      Malignant   56      444
+```
+### Test Set (750 images — held out)
+| Metric | Value |
+|--------|-------|
+| **Accuracy** | 87.2% |
+| **F1-Score** | 90.8% |
+| **Sensitivity (Recall)** | 87.5% |
+| **Specificity** | 86.5% |
+| **PPV (Precision)** | 94.4% |
+| **NPV** | 72.6% |
+| **AUC-ROC** | **0.937** |
+**Confusion Matrix:**
+```
+              Predicted
+           Benign  Malignant
+Actual Benign    180       28
+      Malignant   68      474
+```
+### Training Curves
+The model converged around epoch 18-22 with validation AUC-ROC peaking at 0.940. Early stopping triggered at epoch 27, loading the best checkpoint.
+| Epoch | Train Loss | Val AUC-ROC | Val Sensitivity | Val Specificity |
+|-------|-----------|-------------|-----------------|-----------------|
+| 1 | 0.356 | 0.713 | 0.714 | 0.590 |
+| 5 | 0.229 | 0.912 | 0.940 | 0.715 |
+| 10 | 0.187 | 0.922 | 0.858 | 0.835 |
+| 15 | 0.148 | 0.934 | 0.928 | 0.805 |
+| 18 | 0.125 | **0.939** | 0.846 | 0.885 |
+| 22 | 0.143 | **0.940** | 0.888 | 0.855 |
 ---
+## External Validation
+To assess generalization, we tested the model on an independent dataset without any fine-tuning:
+| Metric | Value |
+|--------|-------|
+| **Accuracy** | 66.8% |
+| **F1-Score** | 44.7% |
+| **Sensitivity** | 34.5% |
+| **Specificity** | 87.4% |
+| **PPV** | 63.5% |
+| **NPV** | 67.7% |
+| **AUC-ROC** | **0.707** |
+**Confusion Matrix (External):**
+```
+              Predicted
+           Benign  Malignant
+Actual Benign   1665      240
+      Malignant  793      417
+```
+**Analysis:** The external validation shows a significant performance drop (AUC 0.94 → 0.71), which is expected due to:
+1. **Domain shift:** Different ultrasound machines, protocols, and image preprocessing
+2. **Different ROI extraction:** The external dataset may use different cropping strategies
+3. **Population differences:** Different patient demographics and disease prevalence
+This highlights the importance of **domain adaptation** or **fine-tuning on local data** before clinical deployment.
+---
+## How to Use
+### Quick Inference with Pipeline
 ```python
+from transformers import pipeline
+from PIL import Image
+# Load model
+classifier = pipeline("image-classification", model="Johnyquest7/TN5000_model")
+# Predict
+image = Image.open("thyroid_ultrasound.png").convert("RGB")
+results = classifier(image)
+# Results format:
+# [{'label': 'malignant', 'score': 0.944}, {'label': 'benign', 'score': 0.056}]
+```
+### Manual Inference
+```python
+import torch
+from PIL import Image
+from transformers import AutoImageProcessor, AutoModelForImageClassification
+# Load model and processor
+model = AutoModelForImageClassification.from_pretrained("Johnyquest7/TN5000_model")
+processor = AutoImageProcessor.from_pretrained("Johnyquest7/TN5000_model")
+# Preprocess
+image = Image.open("thyroid_ultrasound.png").convert("RGB")
+inputs = processor(image, return_tensors="pt")
+# Predict
+with torch.no_grad():
+    logits = model(**inputs).logits
+    probs = torch.softmax(logits, dim=-1)[0]
+# Get probabilities
+malignant_prob = probs[1].item()
+benign_prob = probs[0].item()
+print(f"Malignant: {malignant_prob:.1%}")
+print(f"Benign: {benign_prob:.1%}")
 ```
+### Gradio Demo
+Try the live demo: [🩺 Thyroid Nodule Classifier Demo](https://huggingface.co/spaces/Johnyquest7/tn5000-thyroid-demo)
+---
+## Limitations & Disclaimers
+⚠️ **CRITICAL: This model is for research and educational purposes only.**
+1. **Not FDA-approved** for clinical use
+2. **External validation showed performance degradation** (AUC 0.71 vs 0.94) — domain shift is a real concern
+3. **Trained on ROI patches**, not full ultrasound images — the model expects pre-cropped nodule regions
+4. **Class imbalance** in training data may bias predictions
+5. **No multi-institutional validation** — performance may vary across hospitals and equipment
+6. **Always consult a radiologist or endocrinologist** for diagnosis
+**Intended Use Cases:**
+- Research on AI-assisted thyroid screening
+- Educational tool for medical students
+- Prototype for integration into PACS systems (with proper validation)
+**Not Intended For:**
+- Direct patient diagnosis
+- Replacing human radiologists
+- Screening without supervision
+---
+## Citation
+If you use this model in your research, please cite:
+```bibtex
+@misc{tn5000_model,
+  title={TN5000 Thyroid Ultrasound Classifier},
+  author={Johnyquest7},
+  year={2026},
+  howpublished={\url{https://huggingface.co/Johnyquest7/TN5000_model}},
+  note={Fine-tuned SwinV2-Base for benign vs malignant thyroid nodule classification}
+}
+```
+**Base Model:**
+```bibtex
+@article{liu2022swinv2,
+  title={Swin Transformer V2: Scaling Up Capacity and Resolution},
+  author={Liu, Ze and Hu, Han and Lin, Yutong and Yao, Zhuliang and Xie, Zhenda and Wei, Yixuan and Ning, Jia and Cao, Yue and Zhang, Zheng and Dong, Li and Wei, Furu and Guo, Baining},
+  journal={International Conference on Computer Vision (ICCV)},
+  year={2021}
+}
+```
+**Dataset:**
+- TN5000 ROI Dataset: [Kaggle](https://www.kaggle.com/datasets/sadib2026/roi-dataset-tn5000)
+---
+## Acknowledgments
+- Model trained using Hugging Face Transformers and Datasets libraries
+- Compute provided by Hugging Face GPU credits
+- Base model: Microsoft SwinV2-Base
+---
+*Generated by ML Intern — an agent for machine learning research and development on the Hugging Face Hub.*