Update README.md to emphasize CUDA requirement and remove CPU fallback references

README.md

@@ -14,7 +14,7 @@ The project consists of two main components:
 - CMake (3.18 or higher)
 - C++17 compatible compiler
 - LibTorch (PyTorch C++ API)
-- CUDA (optional, for GPU acceleration)
+- **CUDA (required)** - This implementation requires CUDA and does not support CPU-only execution
 
 ## Building the Project
 
@@ -28,8 +28,8 @@ chmod +x build.sh
 ```
 
 This will:
-1. Check for CUDA availability
-2. Download LibTorch if not already installed
+1. Check for CUDA availability (and exit if not found)
+2. Download LibTorch with CUDA support if not already installed
 3. Configure the project with CMake
 4. Build the project
 5. Install the executable to the `bin/` directory
@@ -50,16 +50,12 @@ cmake --build . --config Release
 To run the demo application:
 
 ```bash
-# Set the library path to include LibTorch
-LD_LIBRARY_PATH=$HOME/libtorch/lib:$LD_LIBRARY_PATH ./bin/tracking_demo
-```
-
-Or use the provided script:
-
-```bash
+# Make sure CUDA is properly set up in your environment
 ./run_demo.sh
 ```
 
+The script will check for CUDA availability and set up the necessary environment variables before running the demo.
+
 ## Project Structure
 
 - `cimp/`: Main C++ implementation
@@ -73,10 +69,11 @@ Or use the provided script:
 - `ltr/`: Reference Python implementation
 - `bin/`: Built executables
 
-## Known Issues
+## Implementation Notes
 
-- The PrRoIPooling implementation requires CUDA, but there's a fallback CPU implementation
-- Some CUDA operations may fail on certain GPU configurations; the code includes fallbacks
+- The PrRoIPooling implementation requires CUDA and has no CPU fallback
+- All tensor operations are performed on CUDA devices
+- The tracker is optimized for GPU execution only
 
 ## Comparing Python and C++ Implementations