added workflow, refactored

Author: vaibhavbahel

.github/workflows/inference-docker.yml

@@ -0,0 +1,51 @@
+name: Build & Push Inference Image
+
+on:
+  push:
+    branches: [main]
+    tags: ["v*"]
+    paths:
+      - "src/baseline-inference/**"
+      - ".github/workflows/inference-docker.yml"
+
+jobs:
+  build-push:
+    runs-on: ubuntu-latest
+    permissions:
+      contents: read
+      packages: write
+
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      - name: Log in to GHCR
+        uses: docker/login-action@v3
+        with:
+          registry: ghcr.io
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Extract metadata
+        id: meta
+        uses: docker/metadata-action@v5
+        with:
+          images: ghcr.io/bic-mac-challenge/inference
+          tags: |
+            type=semver,pattern={{version}}
+            type=semver,pattern={{major}}.{{minor}}
+            type=sha,prefix=sha-,format=short
+            type=raw,value=latest,enable={{is_default_branch}}
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
+      - name: Build and push
+        uses: docker/build-push-action@v6
+        with:
+          context: src/baseline-inference
+          push: true
+          tags: ${{ steps.meta.outputs.tags }}
+          labels: ${{ steps.meta.outputs.labels }}
+          cache-from: type=registry,ref=ghcr.io/bic-mac-challenge/inference:cache
+          cache-to: type=registry,ref=ghcr.io/bic-mac-challenge/inference:cache,mode=max

README.md

@@ -116,6 +116,54 @@ python src/baseline/model.py data/sub-000/features/ results/sub-000/ct_pred.nii.
 ```
 
 ---
+## 🐳 Inference Docker (`src/baseline-inference/`)
+
+To simplify reproducibility and submission, the baseline model is also provided as a **fully self-contained Docker image**. This container wraps the same baseline UNet model and runs inference directly from the command line.
+
+This Docker image should be considered the **official baseline submission** — participants are expected to improve upon it.
+
+### 📦 Pull Image
+
+```bash
+docker pull ghcr.io/bic-mac-challenge/inference:latest
+```
+
+### ▶️ Run Inference
+
+```bash
+docker run --rm \
+  --gpus all \
+  -v /absolute/path/to/data:/data \
+  ghcr.io/bic-mac-challenge/inference:latest \
+  --input /data/sub-XXX/features/nacpet.nii.gz \
+  --output /data/sub-XXX/pseudo_ct.nii.gz
+```
+
+### 📁 Input
+
+* Expects NAC-PET input (`nacpet.nii.gz`)
+* Path must be mounted into the container (e.g., `/data`)
+
+### 📁 Output
+
+* Writes predicted CT (in HU) to the specified output path
+
+### ⚙️ Notes
+
+* All model weights and dependencies are **baked into the container**
+* No internet access is required at runtime
+* Designed to be **fully compliant with challenge submission requirements**
+* Serves as a **reference baseline for participants**
+
+### 🧠 Relation to Baseline Code
+
+This container directly wraps the implementation in:
+
+```
+src/baseline/
+```
+
+with identical preprocessing, model architecture, and inference logic.
 
 ## ⚙️ Reconstruction Pipeline (`src/recon/`)
 

…/baseline/v2/inference/docker/Dockerfile src/baseline-inference/Dockerfilesrc/baseline/v2/inference/docker/Dockerfile renamed to src/baseline-inference/Dockerfile src/baseline/v2/inference/docker/Dockerfile renamed to src/baseline-inference/Dockerfile

@@ -1,17 +1,20 @@
-# 🧠 Pseudo-CT Inference Docker
+# 🧠 Baseline Inference Docker (Pseudo-CT)
 
-This repository provides a **Dockerized inference pipeline** for generating pseudo-CT images from NAC PET input using a trained 3D U-Net model.
+This repository provides a **Dockerized baseline inference pipeline** for generating pseudo-CT images from NAC PET input using a trained 3D U-Net model.
+
+This container corresponds to the **official baseline model** of the BIC-MAC challenge.
 
 ---
 
 ## 🚀 Features
 
 * ✅ GPU-accelerated inference (CUDA)
 * ✅ Sliding window inference (overlap = 0.75)
-* ✅ PET-only input
+* ✅ PET-only input (baseline setting)
 * ✅ Outputs CT in Hounsfield Units (HU)
 * ✅ Clean CLI interface
 * ✅ Runtime + inference timing logs
+* ✅ Fully self-contained (no external downloads)
 
 ---
 
@@ -23,32 +26,32 @@ This repository provides a **Dockerized inference pipeline** for generating pseu
 
 ---
 
-## 🏗️ Build Docker Image
-
-Navigate to the docker directory:
+## 📦 Pull Prebuilt Image (Recommended)
 
 ```bash
-cd inference/docker
+docker pull ghcr.io/bic-mac-challenge/inference:latest
 ```
 
-Build the image:
+---
+
+## 🏗️ Build Docker Image (Optional)
+
+If you want to build locally:
 
 ```bash
-docker build -t pseudoct .
+docker build -t ghcr.io/bic-mac-challenge/inference:latest .
 ```
 
 ---
 
 ## ▶️ Run Inference
 
-### 🔹 Basic Usage
-
 ```bash
 docker run --rm \
   --gpus all \
   --user $(id -u):$(id -g) \
   -v /absolute/path/to/data:/data \
-  pseudoct \
+  ghcr.io/bic-mac-challenge/inference:latest \
   --input /data/sub-XXX/features/nacpet.nii.gz \
   --output /data/sub-XXX/pseudo_ct.nii.gz
 ```
@@ -116,6 +119,14 @@ Otherwise inference will run on CPU.
 
 ---
 
+### 🔸 Self-contained Model
+
+* All model weights are **baked into the container**
+* No internet access is required
+* Suitable for **offline evaluation environments (e.g., Codabench)**
+
+---
+
 ## ⏱️ Output Logs
 
 Example output:
@@ -136,7 +147,7 @@ Saved: /data/sub-XXX/pseudo_ct.nii.gz
 ```bash
 docker run --rm -it \
   -v /your/data:/data \
-  pseudoct bash
+  ghcr.io/bic-mac-challenge/inference:latest bash
 
 ls /data
 ```
@@ -157,9 +168,19 @@ inside `inference.py`.
 
 ## 📌 Summary
 
-| Step  | Command                                            |
-| ----- | -------------------------------------------------- |
-| Build | `docker build -t pseudoct .`                       |
-| Run   | `docker run ... pseudoct --input ... --output ...` |
+| Step | Command                                              |
+| ---- | ---------------------------------------------------- |
+| Pull | `docker pull ghcr.io/bic-mac-challenge/inference`    |
+| Run  | `docker run ... ghcr.io/bic-mac-challenge/inference` |
 
 ---
+
+## 🧠 Relation to Challenge
+
+This container represents the **baseline pseudo-CT model** provided in:
+
+```
+src/baseline/
+```
+
+Participants are expected to **build upon and improve this baseline** using additional modalities such as MRI and topogram.