visualization webgl threejs machine-learning interactive data-visualization 3d paper-review research-papers apple-ml
-
Updated
Apr 6, 2026 - HTML
go-mlx native inference, training and research tooling.
MLX-native port of KVarN — variance-normalized KV-cache quantization for Apple Silicon. 3.3× compression at 71% FP16 speed, matches FP16 on GSM8K within ~2%.
To associate your repository with the apple-ml topic, visit your repo's landing page and select "manage topics."