---
title: BrainAge Training
emoji: 🧠
colorFrom: green
colorTo: blue
sdk: docker
app_port: 7860
pinned: false
license: cc-by-nc-4.0
---

# BrainAge Training Pipeline

GPU-powered training Space for the BrainAge brain-age prediction model.

**Workflow:**
1. Downloads preprocessed `.pt` cache from [BrainAge-Golden-Preprocessed](https://huggingface.co/datasets/bilalahmad176176/BrainAge-Golden-Preprocessed)
2. Generates stratified train/val/test split
3. Phase A: Lifespan pretraining (all ages, 60 epochs)
4. Phase B: Pediatric fine-tuning (0–25y, 30 epochs)
5. Pushes checkpoints and logs to [BrainAge-SFCN](https://huggingface.co/bilalEthizo/BrainAge-SFCN)

**Required secrets:** `HF_TOKEN` (write access to model repo)