ELS-RD / kernl
Kernl lets you run PyTorch transformer models several times faster on GPU with a single line of code, and is designed to be easily hackable.
☆1,560Updated last year
Alternatives and similar repositories for kernl:
Users that are interested in kernl are comparing it to the libraries listed below
- A Python-level JIT compiler designed to make unmodified PyTorch programs faster.☆1,038Updated 11 months ago
- MII makes low-latency and high-throughput inference possible, powered by DeepSpeed.☆1,998Updated last week
- A library for accelerating Transformer models on NVIDIA GPUs, including using 8-bit floating point (FP8) precision on Hopper and Ada GPUs…☆2,311Updated last week
- Code for the ICLR 2023 paper "GPTQ: Accurate Post-training Quantization of Generative Pretrained Transformers".☆2,073Updated last year
- Efficient, scalable and enterprise-grade CPU/GPU inference server for 🤗 Hugging Face transformer models 🚀☆1,683Updated 5 months ago
- [ICML 2023] SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Models☆1,373Updated 8 months ago
- Pipeline Parallelism for PyTorch☆761Updated 7 months ago
- Cramming the training of a (BERT-type) language model into limited compute.☆1,326Updated 9 months ago
- Foundation Architecture for (M)LLMs☆3,067Updated 11 months ago
- Implementation of RETRO, Deepmind's Retrieval based Attention net, in Pytorch☆860Updated last year
- Automatically Discovering Fast Parallelization Strategies for Distributed Deep Neural Network Training☆1,779Updated 2 weeks ago
- Central place for the engineering/scaling WG: documentation, SLURM scripts and logs, compute environment and data.☆990Updated 8 months ago
- Ongoing research training transformer language models at scale, including: BERT & GPT-2☆1,382Updated last year
- Training and serving large-scale neural networks with auto parallelization.☆3,123Updated last year
- AITemplate is a Python framework which renders neural network into high performance CUDA/HIP C++ code. Specialized for FP16 TensorCore (N…☆4,622Updated this week
- An open-source efficient deep learning framework/compiler, written in python.☆692Updated last month
- PyTorch extensions for high performance and large scale training.☆3,285Updated 2 months ago
- Tutel MoE: An Optimized Mixture-of-Experts Implementation☆790Updated this week
- maximal update parametrization (µP)☆1,486Updated 8 months ago
- Library for 8-bit optimizers and quantization routines.☆717Updated 2 years ago
- [ICML 2024] Break the Sequential Dependency of LLM Inference Using Lookahead Decoding☆1,229Updated 3 weeks ago
- Microsoft Automatic Mixed Precision Library☆587Updated 6 months ago
- A pytorch quantization backend for optimum☆910Updated 3 weeks ago
- [ICML 2024] SqueezeLLM: Dense-and-Sparse Quantization☆683Updated 7 months ago
- Tile primitives for speedy kernels☆2,208Updated this week
- Ongoing research training transformer language models at scale, including: BERT & GPT-2☆2,034Updated last week
- Transformer related optimization, including BERT, GPT☆6,100Updated last year
- [MLSys 2024 Best Paper Award] AWQ: Activation-aware Weight Quantization for LLM Compression and Acceleration☆2,895Updated last week
- FP16xINT4 LLM inference kernel that can achieve near-ideal ~4x speedups up to medium batchsizes of 16-32 tokens.☆783Updated 7 months ago
- Serving multiple LoRA finetuned LLM as one☆1,044Updated 10 months ago