eDiff-I

Closed weights NVIDIA 9.1B parameters November 2022

No estimate

No hardware requirements for this model

The weights for this model have not been published, so it cannot be downloaded or run on your own hardware at any size. It is reachable only through its provider, and no graphics card changes that.

On record

Full specification

Everything on record for this model. Most of it describes how it was trained rather than how it runs — useful context for judging how much work went into it, and how it compares with models built at a different scale.

Origin

Who built this model, where, and when it was published.

Organisation
NVIDIA
Organisation type
Industry
Country
United States of America
Published
2 November 2022
Authors
Yogesh Balaji, Seungjun Nah, Xun Huang, Arash Vahdat, Jiaming Song, Qinsheng Zhang, Karsten Kreis, Miika Aittala, Timo Aila, Samuli Laine, Bryan Catanzaro, Tero Karras, Ming-Yu Liu

What it does

The problem areas the model was built for. A model can carry several of each.

Domain
Image generation
Task
Image generation, Text-to-image

Size

How large the model is and how much data it was trained on. Parameters are the figure that decides whether it fits on a given graphics card.

Parameters
9.1B

9.1B for config D, Table 1

Training data
1,556,275,200,000 tokens

"The final dataset to train our model contains about one billion text-image pairs"

Training compute

The arithmetic performed to train the model, measured in floating-point operations. It is a measure of what the training run cost, not of how fast the finished model answers you.

Training compute
5.5 × 10¹⁹ FLOP

6ND = 6*9100000000*1000000000=5.46e+19 (likely, might change because of several epochs / dataset division) "The base model was trained using 256 NVIDIA A100 GPUs, while the two super-resolution models were trained with 128 NVIDIA A100 GPUs each" no info on duration

How it was established
Operation counting

The training run

What it physically took to train: which chips, how many, for how long, and what that drew from the wall.

Training hardware
NVIDIA A100

Availability

Whether you can obtain the model and run it on your own hardware, which is what decides if any of the graphics-card figures on this page apply.

Weights
Closed — provider access only
Model access
Unreleased
Training code
Unreleased

How it is classified

Labels the source dataset applies when tracking notable models, and how confident it is in the entry.

Why it is tracked
SOTA improvement

SOTA zero-shot FID on COCO 2014, Table 1 May be significantly used, via Nvidia Picasso: https://www.nvidia.com/en-us/gpu-cloud/picasso/

Record confidence
Likely
Citations
1,046

Sources

Where this record came from and when it was last checked.

Reference
eDiff-I: Text-to-Image Diffusion Models with an Ensemble of Expert Denoisers
Last updated
25 May 2026

What the numbers mean

About this model

eDiff-I was published by NVIDIA, in United States of America, in November 2022. The organisation is categorised as industry.

It works in Image generation, and is recorded as doing image generation, Text-to-image.

This is a closed model: the trained values stayed with whoever produced them, and there is no local version to run.

Training and provenance

Training it took roughly 5.5 × 10¹⁹ FLOP of computation, on NVIDIA A100 — a measure of what producing the model cost, not of how fast it answers.

Around 1,556,275,200,000 tokens went into training it.

Its inclusion criterion is sOTA improvement.

Answers

eDiff-I — common questions

01

Is eDiff-I open source?

No. eDiff-I has not had its weights published, so it exists only as a service controlled by its owner.

02

How many parameters does eDiff-I have?

eDiff-I has 9.1B parameters. 9.1B for config D, Table 1. That figure is the total, and it is what decides how much memory the model needs — roughly half a gigabyte per billion at the compression most people use.

03

Who created eDiff-I?

eDiff-I was published by NVIDIA, based in United States of America, categorised as industry.

04

When was eDiff-I released?

eDiff-I was published in November 2022. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.

05

What is eDiff-I used for?

eDiff-I works in Image generation, and is recorded as handling image generation, Text-to-image. A model can carry several of each, so these are the areas it was built for rather than a limit on what it will attempt.

06

How much compute was used to train eDiff-I?

Around 5.5 × 10¹⁹ FLOP, on NVIDIA A100. That measures what producing the model cost and says nothing about how quickly it answers once trained — inference speed comes from memory bandwidth, not from the training budget.

07

What GPU do I need to run eDiff-I?

None. eDiff-I is a closed model — its weights were never published, so it cannot be downloaded or run on your own hardware at any price. It is reachable only through its provider.

Source

Original publication

Record last updated 25 May 2026

The other direction

Looking at it from the other side?

This page starts from the model. If you already own a card and want to know everything it will run, start from the hardware instead.