DiscDiff

Closed weights Imperial College London February 2024

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
Imperial College London
Organisation type
Academia
Country
United Kingdom of Great Britain and Northern Ireland
Published
8 February 2024
Authors
Zehui Li, Yuhao Ni, William A V Beardall, Guoxuan Xia, Akashaditya Das, Guy-Bart Stan, Yiren Zhao

What it does

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

Domain
Biology
Task
Protein or nucleotide language model (pLM/nLM)

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.

Training data
983,040,000 tokens

160000 examples with 2048 length (Table 1) 1.6e+5*2.0e+3=3.3e+8

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
3.4 × 10¹⁹ FLOP

1. Hardware setup: - VAE stage: 1x NVIDIA RTX A6000 (3.87e13 FLOP/s) - UNet stage: 1x NVIDIA A100 40GB (3.12e14 FLOP/s) 2. Training duration (provided directly): - VAE: 24 GPU-hours (86,400 seconds) - UNet: 72 GPU-hours (259,200 seconds) 3. Utilization rate: 40% for both stages 4. Calculation: VAE: 3.87e13 FLOP/s × 86,400s × 0.4 = 1.34e18 FLOPs UNet: 3.12e14 FLOP/s × 259,200s × 0.4 = 3.24e19 FLOPs Total = 1.34e18 + 3.24e19 = 3.4e19 FLOPs

How it was established
Hardware

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 RTX A6000,NVIDIA A100
Chips used
2

How it is classified

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

Record confidence
Confident
Citations
21

Sources

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

Reference
DiscDiff: Latent Diffusion Model for DNA Sequence Generation
Last updated
25 May 2026

What the numbers mean

About this model

DiscDiff was published by Imperial College London, in United Kingdom of Great Britain and Northern Ireland, in February 2024. academia is the category the publisher falls under.

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

Its weights were never published, so it can only be reached through its provider. No graphics card changes that.

Training and provenance

Training it took roughly 3.4 × 10¹⁹ FLOP of computation, on NVIDIA RTX A6000,NVIDIA A100 — a measure of what producing the model cost, not of how fast it answers.

It was trained on about 983,040,000 tokens of text.

Answers

DiscDiff — common questions

01

Is DiscDiff open source?

The licensing for DiscDiff 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 DiscDiff have?

No parameter count has been published for DiscDiff, which is why no memory or speed figure appears on this page.

03

Who created DiscDiff?

DiscDiff was published by Imperial College London, based in United Kingdom of Great Britain and Northern Ireland, categorised as academia.

04

When was DiscDiff released?

DiscDiff was published in February 2024. 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 DiscDiff used for?

DiscDiff works in Biology, and is recorded as handling 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

How much compute was used to train DiscDiff?

Around 3.4 × 10¹⁹ FLOP, on NVIDIA RTX A6000,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 DiscDiff?

None. DiscDiff 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.