HazyResearch / aisys-building-blocks
Building blocks for foundation models.
β394Updated 10 months ago
Related projects β
Alternatives and complementary repositories for aisys-building-blocks
- Helpful tools and examples for working with flex-attentionβ469Updated 3 weeks ago
- Implementation of π Ring Attention, from Liu et al. at Berkeley AI, in Pytorchβ476Updated 3 weeks ago
- A subset of PyTorch's neural network modules, written in Python using OpenAI's Triton.β483Updated 3 weeks ago
- Puzzles for learning Tritonβ1,135Updated this week
- A bibliography and survey of the papers surrounding o1β754Updated this week
- Annotated version of the Mamba paperβ457Updated 8 months ago
- What would you do with 1000 H100s...β903Updated 10 months ago
- π Efficiently (pre)training foundation models with native PyTorch features, including FSDP for training and SDPA implementation of Flashβ¦β193Updated this week
- Transformers with Arbitrarily Large Contextβ641Updated 3 months ago
- Efficient implementations of state-of-the-art linear attention models in Pytorch and Tritonβ1,339Updated this week
- Cataloging released Triton kernels.β134Updated 2 months ago
- This repository contains the experimental PyTorch native float8 training UXβ211Updated 3 months ago
- Pipeline Parallelism for PyTorchβ726Updated 2 months ago
- β224Updated 4 months ago
- Small scale distributed training of sequential deep learning models, built on Numpy and MPI.β107Updated last year
- β197Updated 4 months ago
- π Collection of components for development, training, tuning, and inference of foundation models leveraging PyTorch native components.β165Updated this week
- Ring attention implementation with flash attentionβ585Updated last week
- Puzzles for exploring transformersβ325Updated last year
- Applied AI experiments and examples for PyTorchβ166Updated 2 weeks ago
- β133Updated 9 months ago
- An ML Systems Onboarding listβ545Updated this week
- Legible, Scalable, Reproducible Foundation Models with Named Tensors and Jaxβ516Updated this week
- Flash Attention in ~100 lines of CUDA (forward pass only)β626Updated 7 months ago
- β391Updated last month
- Collection of kernels written in Triton languageβ68Updated 3 weeks ago
- A repository for research on medium sized language models.β479Updated this week
- Triton-based implementation of Sparse Mixture of Experts.β185Updated last month
- GPU programming related news and material linksβ1,237Updated last month