RetinaNet-R101

Closed weights Facebook AI Research 53M parameters August 2017

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
Facebook AI Research
Organisation type
Industry
Country
United States of America, France
Published
7 August 2017
Authors
TY Lin, P Goyal, R Girshick, K He, P Dollar

What it does

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

Domain
Vision
Task
Object detection
Numerical format
FP32

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
53M

source: table 2 in https://arxiv.org/pdf/1911.09070.pdf

Training data
115,000 tokens

trainval135k split

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
2.1 × 10¹⁸ FLOP

"We use synchronized SGD over 8 GPUs with a total of 16 images per minibatch (2 images per GPU). Unless otherwise specified, all models are trained for 90k iterations with an initial learning rate of 0.01, which is then divided by 10 at 60k and again at 80k iterations. We use horizontal image flipping as the only form of data augmentation unless otherwise noted. Weight decay of 0.0001 and momentum of 0.9 are used. The training loss is the sum the focal loss and the standard smooth L1 loss used f…

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 M40
Chips used
8
Chip-hours
280
Wall-clock time
35 hours

"We use synchronized SGD over 8 GPUs with a total of 16 images per minibatch (2 images per GPU). Unless otherwise specified, all models are trained for 90k iterations with an initial learning rate of 0.01, which is then divided by 10 at 60k and again at 80k iterations. We use horizontal image flipping as the only form of data augmentation unless otherwise noted. Weight decay of 0.0001 and momentum of 0.9 are used. The training loss is the sum the focal loss and the standard smooth L1 loss used f…

Power draw
4.2 kW

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
Highly cited
Record confidence
Confident
Citations
16,437

Sources

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

Reference
Focal loss for dense object detection
Last updated
28 November 2025

What the numbers mean

Background

RetinaNet-R101 was published by Facebook AI Research, in United States of America, in August 2017. industry is the category the publisher falls under.

It works in Vision, and is recorded as doing object detection.

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

How it was trained

Producing it required around 2.1 × 10¹⁸ FLOP of arithmetic, on NVIDIA M40, which is a statement about the training budget rather than about inference.

It was trained on about 115,000 tokens of text.

It is tracked in the underlying dataset for one reason in particular: highly cited.

Answers

RetinaNet-R101 — common questions

01

When was RetinaNet-R101 released?

RetinaNet-R101 was published in August 2017. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.

02

What is RetinaNet-R101 used for?

RetinaNet-R101 works in Vision, and is recorded as handling object detection. These are the areas it was designed around; they describe intent rather than a hard boundary.

03

How much compute was used to train RetinaNet-R101?

Around 2.1 × 10¹⁸ FLOP, on NVIDIA M40. 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.

04

What GPU do I need to run RetinaNet-R101?

None. RetinaNet-R101 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.

05

Is RetinaNet-R101 open source?

The licensing for RetinaNet-R101 was never recorded in our source data. We treat unstated licensing as closed, because an unrecorded licence is not one to rely on.

06

How many parameters does RetinaNet-R101 have?

RetinaNet-R101 has 53M parameters. source: table 2 in https://arxiv.org/pdf/1911.09070.pdf. 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.

07

Who created RetinaNet-R101?

RetinaNet-R101 was published by Facebook AI Research, based in United States of America, categorised as industry.

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.