"Codexx team on AMD AI League Hackathon"
Team Idea
Build a toolkit that takes a PyTorch training workload and benchmarks it on: NVIDIA CUDA AMD ROCm Different precision modes: FP32 / FP16 / BF16 Different batch sizes GPU memory usage Training throughput Kernel execution time Scaling across GPUs
Team Members
Selz
selz_674
Submission
No submission yet