Slides

• Tree Positional Encodings for Transformers — based on Shiv & Quirk, NeurIPS 2019
• YAML-BERT — an embedding model for Kubernetes manifests (intro)
• DECODE-RAPL — predicting CPU power from OS-level metrics