Fitsum Reda's picture

2

Fitsum Reda

fitsumreda

·

https://fitsumreda.github.io

AI & ML interests

I'm a principal scientist at NVIDIA, where I work on deep generative models and multimodal LLMs for world models and physical AI.

Recent Activity

updated a model about 1 month ago

nvidia/Cosmos-1.0-Tokenizer-CV8x8x8

updated a model about 1 month ago

nvidia/Cosmos-1.0-Tokenizer-DV8x16x16

upvoted an article about 1 month ago

Announcing NVIDIA Cosmos World Foundation Models

View all activity

Organizations

fitsumreda's activity

updated 2 models about 1 month ago

nvidia/Cosmos-1.0-Tokenizer-CV8x8x8

Updated about 1 month ago • 5.33k • 15

nvidia/Cosmos-1.0-Tokenizer-DV8x16x16

Updated about 1 month ago • 1.2k • 14

upvoted an article about 1 month ago

Article

Announcing NVIDIA Cosmos World Foundation Models

By

and 1 other •

Jan 7

• 23

upvoted a collection about 1 month ago

Cosmos

The collection of Cosmos models • 31 items • Updated 26 days ago • 259

updated 3 models about 2 months ago

nvidia/Cosmos-0.1-Tokenizer-DI16x16

Updated Dec 25, 2024 • 334 • 8

nvidia/Cosmos-0.1-Tokenizer-DI8x8

Updated Dec 25, 2024 • 608 • 11

nvidia/Cosmos-0.1-Tokenizer-CI16x16

Updated Dec 25, 2024 • 2.2k • 8

updated a collection about 2 months ago

Cosmos Tokenizer

A suite of image and video tokenizers • 13 items • Updated 26 days ago • 38

authored 4 papers 3 months ago

Image Inpainting for Irregular Holes Using Partial Convolutions

Paper • 1804.07723 • Published Apr 20, 2018

Partial Convolution based Padding

Paper • 1811.11718 • Published Nov 28, 2018

TryOnDiffusion: A Tale of Two UNets

Paper • 2306.08276 • Published Jun 14, 2023 • 72

FILM: Frame Interpolation for Large Motion

Paper • 2202.04901 • Published Feb 10, 2022