NVIDIA / TransformerEngine
A library for accelerating Transformer models on NVIDIA GPUs, including using 8-bit floating point (FP8) precision on Hopper, Ada and Blackwell GPUs, to provide better performance with lower memory utilization in both training and inference.
☆2,393Updated this week
Alternatives and similar repositories for TransformerEngine:
Users that are interested in TransformerEngine are comparing it to the libraries listed below
- FlashInfer: Kernel Library for LLM Serving☆2,788Updated this week
- nvidia-modelopt is a unified library of state-of-the-art model optimization techniques like quantization, pruning, distillation, speculat…☆900Updated last week
- PyTorch native quantization and sparsity for training and inference☆2,015Updated this week
- [ICML 2023] SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Models☆1,399Updated 9 months ago
- Transformer related optimization, including BERT, GPT☆6,147Updated last year
- [MLSys 2024 Best Paper Award] AWQ: Activation-aware Weight Quantization for LLM Compression and Acceleration☆2,984Updated 3 weeks ago
- Pipeline Parallelism for PyTorch☆764Updated 8 months ago
- Tile primitives for speedy kernels☆2,325Updated this week
- Kernl lets you run PyTorch transformer models several times faster on GPU with a single line of code, and is designed to be easily hackab…☆1,566Updated last year
- MII makes low-latency and high-throughput inference possible, powered by DeepSpeed.☆2,009Updated last month
- Ongoing research training transformer language models at scale, including: BERT & GPT-2☆2,064Updated last month
- Code for the ICLR 2023 paper "GPTQ: Accurate Post-training Quantization of Generative Pretrained Transformers".☆2,099Updated last year
- PyTorch extensions for high performance and large scale training.☆3,309Updated last week
- A Python-level JIT compiler designed to make unmodified PyTorch programs faster.☆1,042Updated last year
- Automatically Discovering Fast Parallelization Strategies for Distributed Deep Neural Network Training☆1,789Updated this week
- Microsoft Automatic Mixed Precision Library☆596Updated 7 months ago
- Ongoing research training transformer language models at scale, including: BERT & GPT-2☆1,386Updated last year
- Tutel MoE: Optimized Mixture-of-Experts Library, Support DeepSeek FP8/FP4☆818Updated this week
- Minimalistic large language model 3D-parallelism training☆1,836Updated this week
- Mirage: Automatically Generating Fast GPU Kernels without Programming in Triton/CUDA☆820Updated this week
- Domain-specific language designed to streamline the development of high-performance GPU/CPU/Accelerators kernels☆1,089Updated this week
- FP16xINT4 LLM inference kernel that can achieve near-ideal ~4x speedups up to medium batchsizes of 16-32 tokens.☆812Updated 8 months ago
- 🚀 Efficient implementations of state-of-the-art linear attention models in Torch and Triton☆2,357Updated this week
- [ICML 2024] Break the Sequential Dependency of LLM Inference Using Lookahead Decoding☆1,243Updated 2 months ago
- Puzzles for learning Triton☆1,614Updated 5 months ago
- SOTA low-bit LLM quantization (INT8/FP8/INT4/FP4/NF4) & sparsity; leading model compression techniques on TensorFlow, PyTorch, and ONNX R…☆2,387Updated last week
- A pytorch quantization backend for optimum☆928Updated 2 weeks ago
- AutoAWQ implements the AWQ algorithm for 4-bit quantization with a 2x speedup during inference. Documentation:☆2,145Updated this week
- The Triton TensorRT-LLM Backend☆832Updated this week
- depyf is a tool to help you understand and adapt to PyTorch compiler torch.compile.☆657Updated 2 weeks ago