andrewkchan / yalm
Yet Another Language Model: LLM inference in C++/CUDA, no libraries except for I/O
☆247Updated last month
Alternatives and similar repositories for yalm:
Users that are interested in yalm are comparing it to the libraries listed below
- CUDA/Metal accelerated language model inference☆512Updated 2 months ago
- Materials for learning SGLang☆275Updated 2 weeks ago
- A throughput-oriented high-performance serving framework for LLMs☆739Updated 5 months ago
- Fast low-bit matmul kernels in Triton☆238Updated this week
- BitBLAS is a library to support mixed-precision matrix multiplications, especially for quantized LLM deployment.☆523Updated last week
- Flash Attention in ~100 lines of CUDA (forward pass only)☆700Updated last month
- [MLSys'25] QServe: W4A8KV4 Quantization and System Co-design for Efficient LLM Serving; [MLSys'25] LServe: Efficient Long-sequence LLM Se…☆512Updated this week
- Fastest kernels written from scratch☆173Updated this week
- Dynamic Memory Management for Serving LLMs without PagedAttention☆290Updated this week
- ☆197Updated 3 weeks ago
- Inference of Mamba models in pure C☆183Updated 11 months ago
- A low-latency & high-throughput serving engine for LLMs☆312Updated 3 weeks ago
- [ICLR'25] Fast Inference of MoE Models with CPU-GPU Orchestration☆193Updated 3 months ago
- Fast Matrix Multiplications for Lookup Table-Quantized LLMs☆229Updated this week
- KV cache compression for high-throughput LLM inference☆115Updated 2 weeks ago
- Cataloging released Triton kernels.☆168Updated last month
- LLM KV cache compression made easy☆397Updated this week
- Applied AI experiments and examples for PyTorch☆225Updated this week
- [MLSys'24] Atom: Low-bit Quantization for Efficient and Accurate LLM Serving☆295Updated 7 months ago
- FP16xINT4 LLM inference kernel that can achieve near-ideal ~4x speedups up to medium batchsizes of 16-32 tokens.☆729Updated 5 months ago
- An efficient GPU support for LLM inference with x-bit quantization (e.g. FP6,FP5).☆234Updated 3 months ago
- ☆180Updated this week
- scalable and robust tree-based speculative decoding algorithm☆332Updated 3 weeks ago
- Puzzles for learning Triton, play it with minimal environment configuration!☆229Updated 2 months ago
- Repository for the QUIK project, enabling the use of 4bit kernels for generative inference - EMNLP 2024☆175Updated 10 months ago
- Mirage: Automatically Generating Fast GPU Kernels without Programming in Triton/CUDA☆746Updated this week
- A high-throughput and memory-efficient inference and serving engine for LLMs☆259Updated 4 months ago
- ☆67Updated 2 months ago