GitHub repository angelos-p/llm-from-scratch offers a hands-on workshop for building a GPT training pipeline from scratch. The workshop focuses on reproducing the GPT-2 model with 124 million parameters using PyTorch.

The PyPI package 'lightning', versions 2.6.2 and 2.6.3, was compromised in a supply chain attack, affecting users of the PyTorch Lightning AI training library. The malicious versions include a hidden _runtime directory containing obfuscated JavaScript that activates upon running pip install lightning.

TorchTPU enables running PyTorch natively on Google’s Tensor Processing Units (TPUs), enhancing performance and hardware portability for large-scale machine learning models. It addresses the challenges of distributed systems by supporting clusters of up to 100,000 chips.