Eve
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
- Harvard Medical School,University of Oxford
- Organisation type
- Academia,Academia
- Country
- United States of America, United Kingdom of Great Britain and Northern Ireland
- Published
- 27 October 2021
- Authors
- Jonathan Frazer, Pascal Notin, Mafalda Dias, Aidan Gomez, Joseph K. Min, Kelly Brock, Yarin Gal and Debora S. Marks
What it does
The problem areas the model was built for. A model can carry several of each.
- Domain
- Biology
- Task
- Protein pathogenicity prediction, Proteins
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
- 15M
- Training data
- 24,167,054,120 tokens
"The Bayesian VAE architecture in EVE is comprised of a symmetric 3-layer encoder & decoder architecture (with 2,000-1,000-300 and 300-1,000-2,000 units respectively) and a latent space of dimension 50 [...] We use a single set of parameters for the encoder (ϕp) and learn a fully-factorized gaussian distribution over the weights of the decoder (θp)" They train a new VAE for each protein, and it doesn't seem like they trim the input sequence length, so the largest model will be the one trained fo…
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
- Open source
The model code is available at https://github.com/OATML-Markslab/EVE MIT license
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
- Record confidence
- Likely
- Citations
- 631
"Our model EVE (evolutionary model of variant effect) not only outperforms computational approaches that rely on labelled data but also performs on par with, if not better than, predictions from high-throughput experiments, which are increasingly used as evidence for variant classification" [Abstract] - SOTA improvement for very specific task
Sources
Where this record came from and when it was last checked.
- Reference
- Disease variant prediction with deep generative models of evolutionary data
- Last updated
- 1 January 2026
What the numbers mean
Where it came from
Eve was published by Harvard Medical School,University of Oxford, in United States of America, in October 2021. It comes out of academia,Academia.
It works in Biology, and is recorded as doing protein pathogenicity prediction, Proteins.
Because the weights are not available, none of the hardware figures elsewhere on this site apply to it.
Training and provenance
The training set ran to roughly 24,167,054,120 tokens.
Its inclusion criterion is sOTA improvement.
Answers
Eve — common questions
What GPU do I need to run Eve?
None. Eve 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.
Is Eve open source?
No. Eve has not had its weights published, so it exists only as a service controlled by its owner.
How many parameters does Eve have?
Eve has 15M parameters. "The Bayesian VAE architecture in EVE is comprised of a symmetric 3-layer encoder & decoder architecture (with 2,000-1,000-300 and 300-1,000-2,000 units respectively) and a latent space of dimension 50 [...] We use a single set of parameters for the encoder (ϕp) and learn a fully-factorized gaussian distribution over the weights of the decoder (θp)" They train a new VAE for each protein, and it doesn't seem like they trim the input sequence length, so the largest model will be the one trained for the largest input protein. Supplementary materials 1 gives statistics for each protein; the longest is 5202, which would indicate a network of size 15,010,300. 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.
Who created Eve?
Eve was published by Harvard Medical School,University of Oxford, based in United States of America, categorised as academia,Academia.
When was Eve released?
Eve was published in October 2021. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.
What is Eve used for?
Eve works in Biology, and is recorded as handling protein pathogenicity prediction, Proteins. These are the areas it was designed around; they describe intent rather than a hard boundary.
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.