TransAct

Closed weights Pinterest 92M parameters May 2023

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
Pinterest
Organisation type
Industry
Country
United States of America
Published
31 May 2023
Authors
Xue Xia, Pong Eksombatchai, Nikil Pancha, Dhruvil Deven Badani, Po-Wei Wang, Neng Gu, Saurabh Vishwas Joshi, Nazanin Farahpour, Zhiyuan Zhang, Andrew Zhai

What it does

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

Domain
Recommendation
Task
Recommender system

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
92M

92M (Table 9)

Training data
tokens

[tokens] "Our training dataset contains 3 billion training instances of 177 million users and 720 million pins." "sequence length is |𝑆 | = 100 " "we choose to include each user’s most recent 100 actions in the sequence. For users with less than 100 actions, we pad the feature to the length of 100 with 0s" "The batch size is 12000" 3B training instances * 100 = 3*10^11 training tokens

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
1.7 × 10²⁰ FLOP

6 FLOP / token / parameter * 92 * 10^6 parameters * 3 * 10^11 tokens [see dataset size notes] = 1.656e+20 FLOP

How it was established
Operation counting

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

Apache 2.0 https://github.com/pinterest/transformer_user_action

How it is classified

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

Record confidence
Likely

Sources

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

Reference
TransAct: Transformer-based Realtime User Action Model for Recommendation at Pinterest
Last updated
28 November 2025

What the numbers mean

What this model is

TransAct was published by Pinterest, in United States of America, in May 2023. It comes out of industry.

It works in Recommendation, and is recorded as doing recommender system.

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

Training it took roughly 1.7 × 10²⁰ FLOP of computation — a measure of what producing the model cost, not of how fast it answers.

Answers

TransAct — common questions

01

Is TransAct open source?

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

02

How many parameters does TransAct have?

TransAct has 92M parameters. 92M (Table 9). 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 TransAct?

TransAct was published by Pinterest, based in United States of America, categorised as industry.

04

When was TransAct released?

TransAct was published in May 2023. 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 TransAct used for?

TransAct works in Recommendation, and is recorded as handling recommender system. These are the areas it was designed around; they describe intent rather than a hard boundary.

06

How much compute was used to train TransAct?

Around 1.7 × 10²⁰ FLOP. 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 TransAct?

None. TransAct 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.