Allowing big data in ML - __int64 changes (Feature #256)
Description
I'm traying to train an AdaBoost with a very big matrix (30000 samples of 80000 dimensions). I have compiled the OpenCV in 64 bits. But, in many places of the code the types are int (Cxcore, ML...). So, it is very small and it doesn't allow to use so big matrix. There is need of change many of the int types to +int64.
Is it expected to be changed?
Related issues
related to Patch #2004: Machine Learning, Problem with large datasets | Done | 2012-05-31 |
Associated revisions
Merge pull request #256 from znah:2.4
Merge pull request #256 from vpisarev:calib3d_fixes
History
Updated by anonymous - almost 15 years ago
See opencv/apps/traincascade for an example on how to use existing MLL Adaboost to train classifiers on a huge out-of-memory amount of data
- Status deleted (
Open)
Updated by Alexander Shishkov about 13 years ago
- Description changed from I'm traying to train an [[AdaBoost]] with a very big matrix (30000 sample... to I'm traying to train an AdaBoost with a very big matrix (30000 samples of... More
Updated by Alexander Shishkov almost 13 years ago
- Status set to Open
Updated by Vadim Pisarevsky almost 13 years ago
let's lower priority of the task with no particular target version set. Normally, when there is a huge amount of training data, it's always possible to subsample it in some way, so it's not a critical feature without which one can not proceed.
- Priority changed from High to Low
Updated by Vadim Pisarevsky about 12 years ago
- Target version deleted ()
- Assignee changed from Vadim Pisarevsky to Leonid Beynenson
Updated by Vadim Pisarevsky about 12 years ago
- Pull request set to https://github.com/Itseez/opencv/pull/32
Updated by Vadim Pisarevsky about 12 years ago
see #2004 and the pull request https://github.com/Itseez/opencv/pull/395 that solves the problem
- Status changed from Open to Cancelled
Updated by Kirill Kornyakov about 12 years ago
- Target version set to 2.4.4