HazyResearch

Tile primitives for speedy kernels

DeepDive

NCM: Next Concept Modeling

Big & Small LLMs working together

Resources for Data Centric AI

Convolutions for Sequence Modeling

Kernels, of the mega variety :)

Explore and understand your training and validation data.

Official implementation for HyenaDNA, a long-range genomic foundation model built with Hyena

Hyperbolic Graph Convolutional Networks in PyTorch.

An open science effort to benchmark legal reasoning in foundation models

Building blocks for foundation models.

Repo for "Monarch Mixer: A Simple Sub-Quadratic GEMM-Based Architecture"

Ask Me Anything language model prompting

Language Modeling with the H3 State Space Model

This repo contains data and code for the paper "Language Models Enable Simple Systems for Generating Structured Views of Heterogeneous Data Lakes"

Fast and Furious AMD Kernels

A tool for converting PDF into hOCR with text, tables, and figures being recognized and preserved.

Prompt programming with FMs.

Snorkel MeTaL: A framework for training models with multi-task weak supervision

A knowledge base construction engine for richly formatted data

Hyperbolic Embeddings

FlashFFTConv: Efficient Convolutions for Long Sequences with Tensor Cores

Storing long contexts in tiny caches with self-study

More interactive weak supervision with FlyingSquid

Understand and test language model architectures on synthetic tasks.

Repo for "LoLCATs: On Low-Rank Linearizing of Large Language Models"

Code for exploring Based models from "Simple linear attention language models balance the recall-throughput tradeoff"

Hyperbolic Knowledge Graph embeddings.

Self-Supervision for Named Entity Disambiguation at the Tail

Hyperbolic Hierarchical Clustering.

TART: A plug-and-play Transformer module for task-agnostic reasoning

Butterfly matrix multiplication in PyTorch

Code for SpaceTime 🌌⏱️. Proposed in Effectively Modeling Time Series with Simple Discrete State Spaces, ICLR 2023.

Learning to Compose Domain-Specific Transformations for Data Augmentation

A system for generating training labels via natural language explanations

Your worst case is our best case.

Automating enterprise workflows with multimodal agents

Tools for iterative knowledge base development with DeepDive

Foundation Models for Data Tasks

Benchmarking Intelligence Efficiency of LM Inference

Automatically labeling training data

A collection of simple tutorials for using Fonduer

Table Extraction Tool

Hyperbolic PCA via Horospherical Projections

Interactive Model Iteration with Weak Supervision and Pre-Trained Embeddings

Structured matrices for compressing neural networks

Combating hidden stratification with GEORGE