Pinned Loading

1. delta_compression_framework delta_compression_framework Public

   A delta compression framework. Implement NTransform, Finesse, Odess, Palantir.

   C 7 5

2. warp_decode warp_decode Public

   An unofficial re-implementation of Cursor's warp decode MoE inference technique. Achieved nearly 2.0x speedup on MOE layer on Nvidia H20 at batch size 1.

   Python 5 1

3. mini-sglang-mlx mini-sglang-mlx Public

   MLX port of mini-sglang. Support Qwen3.8-27B with paged attention, continous batching, prefix caching and DFlash2 speculative decoding.

   Python