Jules Belveze
Resources
Browse the writing by topic. Select a tag to narrow the list.
-
LayerSkip: Early Exiting Grows up for LLMs
Why do decoder-only language models still run every token through every layer?
-
BERxiT: Early Exiting Beyond Entropy
Why should BERT trust its own confidence scores to decide when to stop thinking?
-
Instruction Fine-Tuning Evaluation and Advanced Techniques (opens in a new tab)
You fine-tuned your model to follow instructions—but how do you actually know it works? This guide unpacks the evaluation frameworks and parameter-efficient methods that separate production-ready agents from expensive science projects.
-
DeeBERT: Teaching BERT When to Stop Thinking
Why does BERT need twelve layers to classify “I love this movie” as positive?
-
Early Exiting: The Under-Hyped Compression Method
Why are we burning GPU hours to answer “2 + 2 = 4”?
-
Instruction Fine-Tuning Fundamentals (opens in a new tab)
Instruction Fine-Tuning (IFT) is the secret sauce that transforms generic language models into obedient AI assistants.
-
Scaling Machine Learning Experiments With neptune.ai and Kubernetes (opens in a new tab)
Link post: managing and scaling experiment tracking with Neptune and Kubernetes.
-
Case Study: MLOps for NLP-powered Media Intelligence using Metaflow (opens in a new tab)
Link post: case study on building NLP media intelligence with Metaflow.
-
Scaling-up PyTorch inference: Serving billions of daily NLP inferences with ONNX Runtime (opens in a new tab)
Link post: engineering write-up on large-scale PyTorch inference with ONNX Runtime.
-
Atlastic Reputation AI: Four Years of Advancing and Applying a SOTA NLP Classifier (opens in a new tab)
Link post: paper on advancing and applying a SOTA NLP classifier.