Pinned Loading

1. SlimInfer SlimInfer Public

   [AAAI 2026] SlimInfer: Accelerating Long-Context LLM Inference via Dynamic Token Pruning

   Python 13 1

2. Focus-dLLM Focus-dLLM Public

   [ACL 2026] Focus-dLLM: Accelerating Long-Context Diffusion LLM Inference via Confidence-Guided Context Focusing".

   Python 14

3. inclusionAI/cuLA inclusionAI/cuLA Public

   CUDA kernels for linear attention variants, written in CuTe DSL and CUTLASS C++.

   Python 561 68

4. inclusionAI/ASpec inclusionAI/ASpec Public

   A high-throughput training framework for Speculative decoding models.

   Python 53 6