I am new to NVIDIA Clara and before installing it, I would like to know what the requirements for memory and storage are. In the docs I only found requirements for hardware type. Assuming all requirements are fulfilled (like described here: Installation — Clara Train SDK v3.1 documentation) could you estimate how much memory and storage I will need?
Thanks for your interest in Clara Train SDK. Please note we have recently release clara train V4.0 based on MONAI which uses PyTorch. Please check out the notebooks to get you started clara-train-examples/PyTorch/NoteBooks at master · NVIDIA/clara-train-examples · GitHub
To your question, the requirements varies depending on the workload you will run:
- For AIAA you would be ok with 8GB GPU, 16 or 32 GB main memory.
- For Training you would need more memory 64GB to load data and run preprocessing. You would need to scale this if you have more than 1 GPU.
For storage We managed to get clara train sdk working on AWS with 80GB drive that runs OS and all docker images needed. You definitely need more space for your training data
We are currently working on adding documentation to minimal AWS instances that would run different workloads.
Hope this helps