IST-DASLab / MoE-QuantLinks
Code for data-aware compression of DeepSeek models
☆31Updated last month
Alternatives and similar repositories for MoE-Quant
Users that are interested in MoE-Quant are comparing it to the libraries listed below
Sorting:
- Boosting 4-bit inference kernels with 2:4 Sparsity☆75Updated 9 months ago
- Work in progress.☆67Updated this week
- ☆129Updated 3 months ago
- [ICLR2025] Breaking Throughput-Latency Trade-off for Long Sequences with Speculative Decoding☆116Updated 6 months ago
- ☆130Updated 2 months ago
- GEAR: An Efficient KV Cache Compression Recipefor Near-Lossless Generative Inference of LLM☆163Updated 10 months ago
- 16-fold memory access reduction with nearly no loss☆94Updated 2 months ago
- Code for paper: [ICLR2025 Oral] FlexPrefill: A Context-Aware Sparse Attention Mechanism for Efficient Long-Sequence Inference☆107Updated 2 weeks ago
- KV cache compression for high-throughput LLM inference☆129Updated 3 months ago
- Repository for Sparse Finetuning of LLMs via modified version of the MosaicML llmfoundry☆42Updated last year
- ☆73Updated 4 months ago
- [ICML 2025] SliM-LLM: Salience-Driven Mixed-Precision Quantization for Large Language Models☆31Updated 9 months ago
- Tritonbench is a collection of PyTorch custom operators with example inputs to measure their performance.☆127Updated this week
- Flash-Muon: An Efficient Implementation of Muon Optimizer☆115Updated this week
- DeeperGEMM: crazy optimized version☆69Updated 3 weeks ago
- ☆46Updated last week
- QUICK: Quantization-aware Interleaving and Conflict-free Kernel for efficient LLM inference☆118Updated last year
- Cold Compress is a hackable, lightweight, and open-source toolkit for creating and benchmarking cache compression methods built on top of…☆132Updated 9 months ago
- A minimal implementation of vllm.☆41Updated 10 months ago
- Training-free Post-training Efficient Sub-quadratic Complexity Attention. Implemented with OpenAI Triton.☆132Updated this week
- PyTorch bindings for CUTLASS grouped GEMM.☆93Updated this week
- [NeurIPS 2024] The official implementation of "Kangaroo: Lossless Self-Speculative Decoding for Accelerating LLMs via Double Early Exitin…☆57Updated 11 months ago
- extensible collectives library in triton☆87Updated 2 months ago
- An algorithm for weight-activation quantization (W4A4, W4A8) of LLMs, supporting both static and dynamic quantization☆136Updated last week
- ☆45Updated last year
- ☆49Updated 10 months ago
- ☆99Updated this week
- ShiftAddLLM: Accelerating Pretrained LLMs via Post-Training Multiplication-Less Reparameterization☆107Updated 7 months ago
- ☆59Updated last month
- QuIP quantization☆52Updated last year