IST-DASLab / Sparse-MarlinLinks
Boosting 4-bit inference kernels with 2:4 Sparsity
☆86Updated last year
Alternatives and similar repositories for Sparse-Marlin
Users that are interested in Sparse-Marlin are comparing it to the libraries listed below
Sorting:
- PyTorch bindings for CUTLASS grouped GEMM.☆131Updated 6 months ago
- [ICLR2025] Breaking Throughput-Latency Trade-off for Long Sequences with Speculative Decoding☆132Updated 11 months ago
- ☆151Updated 9 months ago
- Triton-based implementation of Sparse Mixture of Experts.☆253Updated 2 months ago
- QUICK: Quantization-aware Interleaving and Conflict-free Kernel for efficient LLM inference☆118Updated last year
- GEAR: An Efficient KV Cache Compression Recipefor Near-Lossless Generative Inference of LLM☆170Updated last year
- [MLSys'24] Atom: Low-bit Quantization for Efficient and Accurate LLM Serving☆331Updated last year
- ☆113Updated last year
- An efficient GPU support for LLM inference with x-bit quantization (e.g. FP6,FP5).☆272Updated 4 months ago
- Official repository for DistFlashAttn: Distributed Memory-efficient Attention for Long-context LLMs Training☆218Updated last year
- extensible collectives library in triton☆91Updated 8 months ago
- High-speed GEMV kernels, at most 2.7x speedup compared to pytorch baseline.☆123Updated last year
- KV cache compression for high-throughput LLM inference☆143Updated 9 months ago
- ☆83Updated 10 months ago
- Triton-based Symmetric Memory operators and examples☆63Updated last month
- [NeurIPS 2024] KVQuant: Towards 10 Million Context Length LLM Inference with KV Cache Quantization☆391Updated last year
- The official implementation of the EMNLP 2023 paper LLM-FP4☆218Updated last year
- 🚀 Collection of components for development, training, tuning, and inference of foundation models leveraging PyTorch native components.☆217Updated last week
- QuTLASS: CUTLASS-Powered Quantized BLAS for Deep Learning☆143Updated 3 weeks ago
- 16-fold memory access reduction with nearly no loss☆108Updated 8 months ago
- ☆94Updated last year
- Cold Compress is a hackable, lightweight, and open-source toolkit for creating and benchmarking cache compression methods built on top of…☆146Updated last year
- Fast low-bit matmul kernels in Triton☆401Updated last week
- Applied AI experiments and examples for PyTorch☆307Updated 3 months ago
- Framework to reduce autotune overhead to zero for well known deployments.☆88Updated 2 months ago
- [ICLR'25] Fast Inference of MoE Models with CPU-GPU Orchestration☆243Updated last year
- Odysseus: Playground of LLM Sequence Parallelism☆78Updated last year
- Flash-LLM: Enabling Cost-Effective and Highly-Efficient Large Generative Model Inference with Unstructured Sparsity☆224Updated 2 years ago
- ☆132Updated 6 months ago
- Official implementation for Training LLMs with MXFP4☆110Updated 7 months ago