Vespa

Closed weights Technical University of Munich 231K parameters December 2021

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
Technical University of Munich
Organisation type
Academia
Country
Germany
Published
30 December 2021
Authors
Céline Marquet, Michael Heinzinger, Tobias Olenyi, Christian Dallago, Kyra Erckert, Michael Bernhofer, Dmitrii Nechaev & Burkhard Rost

What it does

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

Domain
Biology
Task
Proteins, Protein or nucleotide language model (pLM/nLM)
Approach
Supervised
Base model
ProteinBERT

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
231K

"(3) standard convolutional neural network (CNN; with two convolutional layers with a window size of 7, connected through ReLU activations; dropout rate of 0.25; 231 k free parameters)"

Training data
tokens

ConSurf10k: 2.7 x 10^6 samples Eff10k: 1.00737 x 10^5 samples Total = 2.7 x 10^6 + 1.00737 x 10^5 = 2.800737 x 10^6 ≈ 2.8 x 10^6 datapoints

How it is classified

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

Record confidence
Confident

Sources

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

Reference
Embeddings from protein language models predict conservation and variant effects
Last updated
28 November 2025

What the numbers mean

About this model

Vespa was published by Technical University of Munich, in Germany, in December 2021. It comes out of academia.

It works in Biology, and is recorded as doing proteins, Protein or nucleotide language model (pLM/nLM).

It builds on ProteinBERT, which is why it shares that model's general shape and size.

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

Answers

Vespa — common questions

01

Is Vespa open source?

The licensing for Vespa was never recorded in our source data. We treat unstated licensing as closed, because an unrecorded licence is not one to rely on.

02

How many parameters does Vespa have?

Vespa has 231K parameters. "(3) standard convolutional neural network (CNN; with two convolutional layers with a window size of 7, connected through ReLU activations; dropout rate of 0.25; 231 k free parameters)". 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 Vespa?

Vespa was published by Technical University of Munich, based in Germany, categorised as academia.

04

When was Vespa released?

Vespa was published in December 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.

05

What is Vespa used for?

Vespa works in Biology, and is recorded as handling proteins, Protein or nucleotide language model (pLM/nLM). Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.

06

What GPU do I need to run Vespa?

None. Vespa 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 28 November 2025

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.