Natalie Kershaw articles

Accelerate PyTorch training with torch-ort 

With a simple change to your PyTorch training script, you can now speed up training large language models with torch_ort.ORTModule, running on the target hardware of your choice. Training deep learning models requires ever-increasing compute and memory resources. Today we release torch_ort.