Rational DQN Average

Closed weights TU Darmstadt 1.7M parameters February 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
TU Darmstadt
Organisation type
Academia
Country
Germany
Published
18 February 2021
Authors
Q Delfosse, P Schramowski, A Molina

What it does

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

Domain
Games
Task
Atari

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
1.7M

See figure 7

Training data
tokens

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 (non-commercial)

"To get the trained agents, please contact Quentin Delfosse" no clear license https://github.com/ml-research/rational_rl/

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

They don't claim absolute SOTA "We demonstrate that equipping popular algorithms with (joint) rational activations leads to consistent improvements on Atari games, notably making DQN competitive to DDQN and Rainbow"

Citations
26

Sources

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

Reference
Recurrent Rational Networks
Last updated
25 May 2026

What the numbers mean

Where it came from

Rational DQN Average was published by TU Darmstadt, in Germany, in February 2021. It comes out of academia.

It works in Games, and is recorded as doing atari.

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

What went into building it

It is tracked in the underlying dataset for one reason in particular: sOTA improvement.

Answers

Rational DQN Average — common questions

01

Is Rational DQN Average open source?

No. Rational DQN Average has not had its weights published, so it exists only as a service controlled by its owner.

02

How many parameters does Rational DQN Average have?

Rational DQN Average has 1.7M parameters. See figure 7. 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 Rational DQN Average?

Rational DQN Average was published by TU Darmstadt, based in Germany, categorised as academia.

04

When was Rational DQN Average released?

Rational DQN Average was published in February 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 Rational DQN Average used for?

Rational DQN Average works in Games, and is recorded as handling atari. 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 Rational DQN Average?

None. Rational DQN Average 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.