3DDFA

Closed weights Chinese Academy of Sciences,Michigan State University 5.4M parameters November 2015

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
Chinese Academy of Sciences,Michigan State University
Organisation type
Academia,Academia
Country
China, United States of America
Published
23 November 2015
Authors
Xiangyu Zhu, Zhen Lei, Xiaoming Liu, Hailin Shi, S. Li

What it does

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

Domain
Vision
Task
Face detection, 3D reconstruction
Approach
Supervised

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
5.4M

Parameters based on Figure 2: CNN layers: 4*4*16*6+4*4*32*16=9728 Local CNN: 3*3*48*32*11*11+3*3*64*49*9*9=3958848 FC: 9*9*64*256+234*256=1387008 Total: 9728+3958848+1387008=5355584

Training data
293,901 tokens

"We divide 300W-LP into 97,967 samples for training and 24,483 samples for testing, without identity overlapping. "

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

MIT license: https://github.com/cleardusk/3DDFA (refers to the improved later version of the paper)

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

https://paperswithcode.com/sota/3d-face-reconstruction-on-florence "Experiments on the challenging AFLW database show that our approach achieves significant improvements over state-of-the-art methods."

Record confidence
Confident

Sources

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

Reference
Face Alignment in Full Pose Range: A 3D Total Solution
Last updated
28 November 2025

What the numbers mean

What this model is

3DDFA was published by Chinese Academy of Sciences,Michigan State University, in China, in November 2015. It comes out of academia,Academia.

It works in Vision, and is recorded as doing face detection, 3D reconstruction.

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

Training and provenance

Around 293,901 tokens went into training it.

Its inclusion criterion is sOTA improvement.

Answers

3DDFA — common questions

01

When was 3DDFA released?

3DDFA was published in November 2015. 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 3DDFA used for?

3DDFA works in Vision, and is recorded as handling face detection, 3D reconstruction. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.

03

What GPU do I need to run 3DDFA?

None. 3DDFA 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.

04

Is 3DDFA open source?

No. 3DDFA has not had its weights published, so it exists only as a service controlled by its owner.

05

How many parameters does 3DDFA have?

3DDFA has 5.4M parameters. Parameters based on Figure 2: CNN layers: 4*4*16*6+4*4*32*16=9728 Local CNN: 3*3*48*32*11*11+3*3*64*49*9*9=3958848 FC: 9*9*64*256+234*256=1387008 Total: 9728+3958848+1387008=5355584. 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.

06

Who created 3DDFA?

3DDFA was published by Chinese Academy of Sciences,Michigan State University, based in China, categorised as academia,Academia.

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.