Assignment 2

Infra

What changes when a correct training loop needs to become an efficient one.

• Profiling compute and memory
• GPU kernels and attention
• Distributed training

Stanford assignment materials

← All tracks