Decision tree (classification)
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
- Mitsubishi Electric Research Labs,Compaq CRL
- Organisation type
- Industry,Industry
- Country
- United States of America
- Published
- 8 December 2001
- Authors
- P. Viola, M. Jones
What it does
The problem areas the model was built for. A model can carry several of each.
- Domain
- Vision
- Task
- Face recognition
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
- 12K
- Training data
- 753,616 tokens
Iteratively learns decision tree features, where each feature has 2 parameters (threshold and parity). They learn 6000 features total: "The complete face detection cascade has 38 stages with over 6000 features" Additional weights (see Table 1) are a temporary variable of the optimization and not part of the final model.
Section 5: 4916 hand labeled faces + 9544 non-face images = 14460
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
- 6.3 × 10¹³ FLOP
- How it was established
- Operation counting
The training compute can be tediously worked out from the pseudocode. I think for dataset size D, number of filters T, the training compute is roughly 180k * D * 3 * T = 6.3e13 FLOPs The training evaluates each of 180k candidate features at each step, repeated for 6000 steps (as 1 feature is selected per round). The operations for one feature evaluation are unclear, but should be low (they only compare specific integer positions in the image). Estimated at 10op. 180000*6000*10*14460=1.56168e+…
How it is classified
Labels the source dataset applies when tracking notable models, and how confident it is in the entry.
- Frontier model
- Yes
- Why it is tracked
- Highly cited
- Record confidence
- Likely
- Citations
- 23,449
Sources
Where this record came from and when it was last checked.
- Reference
- Rapid object detection using a boosted cascade of simple features
- Last updated
- 11 February 2026
What the numbers mean
What this model is
Decision tree (classification) was published by Mitsubishi Electric Research Labs,Compaq CRL, in United States of America, in December 2001. It comes out of industry,Industry.
It works in Vision, and is recorded as doing face recognition.
Because the weights are not available, none of the hardware figures elsewhere on this site apply to it.
How it was trained
Producing it required around 6.3 × 10¹³ FLOP of arithmetic, which is a statement about the training budget rather than about inference.
It was trained on about 753,616 tokens of text.
It is tracked in the underlying dataset for one reason in particular: highly cited.
Answers
Decision tree (classification) — common questions
Who created Decision tree (classification)?
Decision tree (classification) was published by Mitsubishi Electric Research Labs,Compaq CRL, based in United States of America, categorised as industry,Industry.
When was Decision tree (classification) released?
Decision tree (classification) was published in December 2001. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.
What is Decision tree (classification) used for?
Decision tree (classification) works in Vision, and is recorded as handling face recognition. A model can carry several of each, so these are the areas it was built for rather than a limit on what it will attempt.
How much compute was used to train Decision tree (classification)?
Around 6.3 × 10¹³ 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.
What GPU do I need to run Decision tree (classification)?
None. Decision tree (classification) 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.
Is Decision tree (classification) open source?
The licensing for Decision tree (classification) was never recorded in our source data. We treat unstated licensing as closed, because an unrecorded licence is not one to rely on.
How many parameters does Decision tree (classification) have?
Decision tree (classification) has 12K parameters. Iteratively learns decision tree features, where each feature has 2 parameters (threshold and parity). They learn 6000 features total: "The complete face detection cascade has 38 stages with over 6000 features" Additional weights (see Table 1) are a temporary variable of the optimization and not part of the final model. 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.
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.