mlfoundations

open_clip

An open source implementation of CLIP.

open_flamingo

An open-source framework for training large multimodal models.

dclm

DataComp for Language Models

MINT-1T

MINT-1T: A one trillion token multimodal interleaved dataset.

datacomp

DataComp: In search of the next generation of multimodal datasets

wise-ft

Robust fine-tuning of zero-shot models

open_lm

A repository for research on medium sized language models.

task_vectors

Editing Models with Task Arithmetic

model-soups

Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference time

evalchemy

Automatic evals for LLMs

open-diffusion

Simple large-scale training of stable diffusion with multi-node support.

scaling

Language models scale reliably with over-training and on downstream tasks

patching

Patching open-vocabulary models by interpolating weights

tableshift

A benchmark for distribution shift in tabular data

VisIT-Bench

imagenet-captions

Release of ImageNet-Captions

rtfm

Research on Tabular Foundation Models

clip_quality_not_quantity

dataset2metadata

tabliblib

A Python library for processing and filtering TabLib

spark-commoncrawl

datacomp_site

webdataset-resharder

Efficiently process webdatasets

advancedml-sp23

imagenet-applications-transfer

au21

llm-foundry

LLM training code for MosaicML foundation models

qa_robustness

MixEval

The official evaluation suite and dynamic data release for MixEval.