commit files to HF hub

README.md

@@ -0,0 +1,49 @@
+---
+license: mit
+tags:
+  - bone-age
+  - regression
+  - medical
+  - resnet
+  - pytorch
+  - onnx
+---
+
+# Bone Age Regression Model (ResNet152, 80M+ params)
+
+This model predicts bone age (in months) from hand X-ray images and gender (male/female).
+- **Base:** ResNet152 pretrained on ImageNet, with added gender embedding and extra layers.
+- **Input:** 256x256 PNG image, gender (0=male, 1=female)
+- **Output:** Bone age in months (float).
+
+## Training Data
+
+- [RSNA Bone Age Dataset on Kaggle](https://www.kaggle.com/datasets/kmader/rsna-bone-age)
+- 12,500 hand X-rays, age labels, gender labels.
+
+## Usage
+
+```python
+from PIL import Image
+import torch
+import onnxruntime as ort
+
+# PyTorch inference
+from finetune_resnet_bone_age import BoneAgeResNet, transforms
+model = BoneAgeResNet()
+model.load_state_dict(torch.load('resnet_bone_age_80m.pt'))
+model.eval()
+img = transforms(Image.open('your_image.png').convert('RGB')).unsqueeze(0)
+gender = torch.tensor([0.0])  # 0 for male, 1 for female
+age_pred = model(img, gender)
+
+# ONNX inference
+ort_sess = ort.InferenceSession('resnet_bone_age_80m.onnx')
+outputs = ort_sess.run(None, {
+    "image": img.numpy(),
+    "gender": gender.unsqueeze(0).numpy()
+})
+print("Predicted age:", outputs[0][0])
+Citation
+If you use this model, please cite the RSNA Bone Age dataset and this repository.
+

resnet_bone_age_80m.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:82fb29361d2e84602a7f1de46f30a3b8232ed15418ca160fad1caaf7b1c8a0ef
+size 241926826

resnet_bone_age_80m.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5606e35db57d71dfbb08538399c6c13b666d15bc539e34f6fac99dec729118b7
+size 243074950