EMMA

arena-evaluation

Running the model with the Arena

datasets

Create efficient datasets for training models

experience-hub

Everything that connects the models together to process and return a response

policy

.github

common

Common modules which are used a lot throughout EMMA repositories.

img2dataset

Easily turn large sets of image urls to an image dataset. Can download, resize and package 100M urls in 20h on one machine.

perception

Object Detector and Feature Extractor

profanity-filter

A Python library for detecting and filtering profanity

VL-T5

PyTorch code for "Unifying Vision-and-Language Tasks via Text Generation" (ICML 2021)

language-evaluation

:clipboard: Collection of evaluation code for natural language generation.

scene-graph-benchmark

A fork of Microsoft's scene graph benchmark