☆75Mar 26, 2025Updated last year
Alternatives and similar repositories for DeepSpeed-Kernels
Users that are interested in DeepSpeed-Kernels are comparing it to the libraries listed below. We may earn a commission when you buy through links labeled 'Ad' on this page.
Sorting:
- A standalone GEMM kernel for fp16 activation and quantized weight, extracted from FasterTransformer☆96Jun 21, 2026Updated 2 months ago
- Performance of the C++ interface of flash attention and flash attention v2 in large language model (LLM) inference scenarios.☆46Feb 27, 2025Updated last year
- MII makes low-latency and high-throughput inference possible, powered by DeepSpeed.☆2,113Jun 30, 2025Updated last year
- High-speed GEMV kernels, at most 2.7x speedup compared to pytorch baseline.☆130Jul 13, 2024Updated 2 years ago
- Official resporitory for "IPDPS' 24 QSync: Quantization-Minimized Synchronous Distributed Training Across Hybrid Devices".☆20Feb 23, 2024Updated 2 years ago
- Deploy on Railway without the complexity - Free Credits Offer • AdConnect your repo and Railway handles the rest with instant previews. Quickly provision container image services, databases, and storage volumes.
- TileFusion is an experimental C++ macro kernel template library that elevates the abstraction level in CUDA C for tile processing.☆118Aug 4, 2026Updated last month
- ☆25Mar 15, 2023Updated 3 years ago
- ☆20Sep 28, 2024Updated last year
- Decoding Attention is specially optimized for MHA, MQA, GQA and MLA using CUDA core for the decoding stage of LLM inference.☆48Jun 11, 2025Updated last year
- ☆272Jul 11, 2024Updated 2 years ago
- ☆13Jan 7, 2025Updated last year
- CUDA 12.2 HMM demos☆21Jul 26, 2024Updated 2 years ago
- A low-latency & high-throughput serving engine for LLMs☆525Jan 8, 2026Updated 8 months ago
- FP16xINT4 LLM inference kernel that can achieve near-ideal ~4x speedups up to medium batchsizes of 16-32 tokens.☆1,148Sep 4, 2024Updated 2 years ago
- Deploy on Railway without the complexity - Free Credits Offer • AdConnect your repo and Railway handles the rest with instant previews. Quickly provision container image services, databases, and storage volumes.
- IntLLaMA: A fast and light quantization solution for LLaMA☆19Jul 21, 2023Updated 3 years ago
- ☆135Nov 11, 2024Updated last year
- A prefill & decode disaggregated LLM serving framework with shared GPU memory and fine-grained compute isolation.☆132Dec 25, 2025Updated 8 months ago
- Dynamic Memory Management for Serving LLMs without PagedAttention☆523Aug 24, 2026Updated 3 weeks ago
- GPTQ inference Triton kernel☆322May 18, 2023Updated 3 years ago
- Implementation for IceFormer: Accelerated Inference with Long-Sequence Transformers on CPUs (ICLR 2024).☆25Jun 9, 2026Updated 3 months ago
- ☆27Feb 20, 2024Updated 2 years ago
- ☆107Sep 9, 2024Updated 2 years ago
- [MLSys'24] Atom: Low-bit Quantization for Efficient and Accurate LLM Serving☆347Jul 2, 2024Updated 2 years ago
- Deploy on Railway without the complexity - Free Credits Offer • AdConnect your repo and Railway handles the rest with instant previews. Quickly provision container image services, databases, and storage volumes.
- MSCCL++: A GPU-driven communication stack for scalable AI applications☆556Updated this week
- Perplexity GPU Kernels☆609Nov 7, 2025Updated 10 months ago
- A Triton JIT runtime and ffi provider in C++☆41Sep 9, 2026Updated last week
- ☆11Apr 5, 2021Updated 5 years ago
- extensible collectives library in triton☆97Mar 31, 2025Updated last year
- FA4-based Relative Attention Kernel developed by TML and Colfax☆18Sep 11, 2026Updated last week
- Stateful LLM Serving☆106Mar 11, 2025Updated last year
- Expert Specialization MoE Solution based on CUTLASS☆27Apr 14, 2026Updated 5 months ago
- Torch Distributed Experimental☆117Aug 5, 2024Updated 2 years ago
- Managed Database hosting by DigitalOcean • AdPostgreSQL, MySQL, MongoDB, Kafka, Valkey, and OpenSearch available. Automatically scale up storage and focus on building your apps.
- A lightweight triton-based General Matrix Multiplication (GEMM) library.☆67Jul 21, 2026Updated last month
- ☆118Jul 28, 2026Updated last month
- High-performance GPU kernels written in TIRx.☆102Updated this week
- optimized BERT transformer inference on NVIDIA GPU. https://arxiv.org/abs/2210.03052☆479Mar 15, 2024Updated 2 years ago
- Standalone Flash Attention v2 kernel without libtorch dependency☆113Sep 10, 2024Updated 2 years ago
- Microsoft Collective Communication Library☆399Aug 25, 2026Updated 3 weeks ago
- Depict GPU memory footprint during DNN training of PyTorch☆11Nov 17, 2022Updated 3 years ago