ConceptAI & ML
HellaSwag evaluation
A sentence completion benchmark dataset used to evaluate a language model's commonsense natural language inference capabilities through multiple-choice questions.
Know first
Named as a prerequisite by the lectures that define this. The number is how many say so. A grey one has no page yet.
Distributed data parallelism1Language model1Multiple-choice evaluation tasks1PyTorch data loaders1Softmax probabilities1Tokenization1
Lecturers
1Connections
2The lectures that define it or assume it, and the concepts next to it.