Document local CUDA deployment

Add the GitHub clone, Git LFS, NVIDIA Container Toolkit, and in-container CUDA verification steps to the Space README.

Co-authored-by: Codex <codex@openai.com>

README.md

@@ -103,10 +103,33 @@ Prerequisites:
 
 - Docker Engine with Docker Compose 2.30 or newer
 - a supported NVIDIA GPU and current NVIDIA driver
-- NVIDIA Container Toolkit when using Docker Engine on Linux
+- NVIDIA Container Toolkit configured for Docker Engine on Linux
+- Git LFS for downloading the interface images
 - enough GPU memory for MiniCPM5-1B and Nemotron-Parse v1.2
 
-Start the application:
+Clone the GitHub repository and download its Git LFS assets:
+
+```bash
+git clone https://github.com/kingabzpro/local-notice-check.git
+cd local-notice-check
+
+git lfs install
+git lfs pull
+```
+
+Before building, confirm that NVIDIA Container Toolkit exposes CUDA and the GPU
+inside Docker:
+
+```bash
+docker run --rm --gpus all \
+  pytorch/pytorch:2.9.1-cuda12.8-cudnn9-runtime \
+  python -c "import torch; print(torch.cuda.is_available(), torch.cuda.get_device_name(0))"
+```
+
+The command must print `True` and the NVIDIA GPU name. If it fails, configure
+NVIDIA Container Toolkit before starting the application.
+
+Build and start NoticeCheck:
 
 ```bash
 docker compose up --build