(reference:launch)=
# Launch PtyRAD
> 💡 This video was recorded for **Ptychography Summer School at Cornell 2025**, which has additional access to SciServer, and a slightly different demo folder structure.
You can launch *PtyRAD* with many different approaches once it's installed in your Python environment.
For example, you can:
1. Run the tutorial Jupyter notebooks under `ptyrad/notebooks/run_ptyrad.ipynb`
2. Launch with CLI tools from your terminal as simple as `ptyrad run `
3. Use a Slurm job script like `ptyrad/scripts/slurm_run_ptyrad.sub`
For all the launching method, *PtyRAD* supports 2 operation modes:
- reconstruction
- hypertune (hyperparameter tuning)
These operation modes can be executed on CPU, GPU, or even distributed on multiple GPUs for better performance.
**Routine workflow**
1. Acquire experimental or get simulated data
2. Prepare params files
3. Launch *PtyRAD* task
4. Analyze and refine the params file
```{toctree}
:maxdepth: 2
:hidden:
python
cli
slurm
hypertune
multiGPU
```