Pinned Loading

1. LowRankClone LowRankClone Public

   [NeurIPS 2025 Spotlight] A Token is Worth over 1,000 Tokens: Efficient Knowledge Distillation through Low-Rank Clone.

   Python 47 1

2. antgroup/OmniKV antgroup/OmniKV Public

   Dynamic Context Selection for Efficient Long-Context LLMs

   Python 56 5

3. Sparse-vLLM Sparse-vLLM Public

   A sparse-first inference engine (sparsevllm). It also contains DeltaKV compressor training + evaluation tooling (deltakv).

   Python 56 8

4. Uni-X Uni-X Public

   [ICLR 2026] Uni-X: Mitigating Modality Conflict with a Two-End-Separated Architecture for Unified Multimodal Models

   Python 10