Models trained during quantization aware fine-tuning experiments using PyTorch's FSDP.