This is the Robotics Learning Notes for ANU 2024 S2 COMP8650
High SoftIRQ on Azure B-Series VM was observed. Investigation revealed that the CPU credits had been exhausted.
Troubleshooting Azure Linux server auto-reboot issue. The cause is Azure auto-installing patches and rebooting the server.
A strange phenomenon where neural network training starts with high accuracy and gradually decreases. The root cause: not setting shuffle=True.
Setonix is a hybrid CPU-GPU supercomputer housed at the Pawsey Centre in Western Australia. This post documents troubleshooting steps for PyTorch and ROCm errors on Setonix.