Alternatives and similar repositories for bfloat16

Users that are interested in bfloat16 are comparing it to the libraries listed below

• IntelLabs / FP8-Emulation-Toolkit
  PyTorch extension for emulating FP8 data formats on standard FP32 Xeon/GPU hardware.
  ☆110Updated 8 months ago
• IBM / triton-dejavu
  Framework to reduce autotune overhead to zero for well known deployments.
  ☆79Updated 2 weeks ago
• jundaf2 / INT8-Flash-Attention-FMHA-Quantization
  ☆158Updated last year
• meta-pytorch / triton-cpu
  An experimental CPU backend for Triton (https//github.com/openai/triton)
  ☆44Updated 4 months ago
• DeMoriarty / custom_matmul_kernels
  Customized matrix multiplication kernels
  ☆56Updated 3 years ago
• Qualcomm-AI-research / FP8-quantization
  ☆154Updated 2 years ago
• wangsiping97 / FastGEMV
  High-speed GEMV kernels, at most 2.7x speedup compared to pytorch baseline.
  ☆113Updated last year
• jaewoosong / pocketnn
  The official, proof-of-concept C++ implementation of PocketNN.
  ☆34Updated last year
• fastmachinelearning / qonnx
  QONNX: Arbitrary-Precision Quantized Neural Networks in ONNX
  ☆151Updated 2 weeks ago
• ROCm / aotriton
  Ahead of Time (AOT) Triton Math Library
  ☆75Updated last week
• cchan / fp8_mul
  A tiny FP8 multiplication unit written in Verilog. TinyTapeout 2 submission.
  ☆14Updated 2 years ago
• benja263 / Integer-Only-Inference-for-Deep-Learning-in-Native-C
  Converting a deep neural network to integer-only inference in native C via uniform quantization and the fixed-point representation.
  ☆25Updated 3 years ago
• nod-ai / transformer-benchmarks
  benchmarking some transformer deployments
  ☆26Updated 2 years ago
• microsoft / TileFusion
  TileFusion is an experimental C++ macro kernel template library that elevates the abstraction level in CUDA C for tile processing.
  ☆93Updated last month
• ruslangrimov / mnist-minimal-model
  Trying to find out what is the minimal model that can achieve 99% accuracy on MNIST dataset
  ☆26Updated 6 years ago
• open-lm-engine / flash-model-architectures
  A bunch of kernels that might make stuff slower 😉
  ☆56Updated last week
• deepspeedai / DeepSpeed-Kernels
  ☆74Updated 4 months ago
• brightlaboratory / polydl
  ☆12Updated 4 years ago
• gau-nernst / quantized-training
  Explore training for quantized models
  ☆20Updated last month
• UmerHA / triton_util
  Make triton easier
  ☆47Updated last year
• mit-han-lab / tinychat-tutorial
  ☆72Updated 9 months ago
• facebookresearch / MODel_opt
  Memory Optimizations for Deep Learning (ICML 2023)
  ☆102Updated last year
• stanford-futuredata / stk
  ☆108Updated 11 months ago
• microsoft / FractalTensor
  FractalTensor is a programming framework that introduces a novel approach to organizing data in deep neural networks (DNNs) as a list of …
  ☆28Updated 7 months ago
• rejunity / tiny-asic-1_58bit-matrix-mul
  Tiny ASIC implementation for "The Era of 1-bit LLMs All Large Language Models are in 1.58 Bits" matrix multiplication unit
  ☆158Updated last year
• LeiWang1999 / AutoGPTQ.tvm
  GPTQ inference TVM kernel
  ☆40Updated last year
• iree-org / iree-bare-metal-arm
  Example for running IREE in a bare-metal Arm environment.
  ☆38Updated 2 weeks ago
• masahi / torchscript-to-tvm
  ☆69Updated 2 years ago
• srush / triton-autodiff
  Experiment of using Tangent to autodiff triton
  ☆80Updated last year
• triton-lang / kernels
  ☆85Updated 9 months ago