CerebrasResearch / Sparse-IFTLinks
Official repository of Sparse ISO-FLOP Transformations for Maximizing Training Efficiency
☆25Updated last year
Alternatives and similar repositories for Sparse-IFT
Users that are interested in Sparse-IFT are comparing it to the libraries listed below
Sorting:
- CUDA and Triton implementations of Flash Attention with SoftmaxN.☆73Updated last year
- Tree Attention: Topology-aware Decoding for Long-Context Attention on GPU clusters☆129Updated 9 months ago
- Experiment of using Tangent to autodiff triton☆81Updated last year
- Token Omission Via Attention☆128Updated 11 months ago
- Understand and test language model architectures on synthetic tasks.☆225Updated 2 months ago
- ☆58Updated 11 months ago
- The simplest, fastest repository for training/finetuning medium-sized GPTs.☆162Updated 2 months ago
- The simplest implementation of recent Sparse Attention patterns for efficient LLM inference.☆87Updated 2 months ago
- ☆89Updated last year
- A MAD laboratory to improve AI architecture designs 🧪☆129Updated 9 months ago
- ☆102Updated 2 months ago
- ☆83Updated last year
- ☆46Updated last year
- Simple and efficient pytorch-native transformer training and inference (batched)☆78Updated last year
- The source code of our work "Prepacking: A Simple Method for Fast Prefilling and Increased Throughput in Large Language Models" [AISTATS …☆60Updated 11 months ago
- ☆53Updated 10 months ago
- nanoGPT-like codebase for LLM training☆107Updated 4 months ago
- train with kittens!☆62Updated 11 months ago
- Code for NeurIPS 2024 Spotlight: "Scaling Laws and Compute-Optimal Training Beyond Fixed Training Durations"☆84Updated 10 months ago
- A fusion of a linear layer and a cross entropy loss, written for pytorch in triton.☆70Updated last year
- Large scale 4D parallelism pre-training for 🤗 transformers in Mixture of Experts *(still work in progress)*☆87Updated last year
- The evaluation framework for training-free sparse attention in LLMs☆96Updated 3 months ago
- ☆85Updated last year
- Landing repository for the paper "Softpick: No Attention Sink, No Massive Activations with Rectified Softmax"☆84Updated last week
- ☆69Updated last year
- seqax = sequence modeling + JAX☆167Updated 2 months ago
- Deep learning library implemented from scratch in numpy. Mixtral, Mamba, LLaMA, GPT, ResNet, and other experiments.☆52Updated last year
- ☆119Updated last year
- some common Huggingface transformers in maximal update parametrization (µP)☆82Updated 3 years ago
- JORA: JAX Tensor-Parallel LoRA Library (ACL 2024)☆36Updated last year