Model Training & Fine-Tuning agent

PyTorch and TensorFlow training loops, LoRA/PEFT fine-tuning, datasets, and distributed training.