- I am using dlib for object detection.I am training my own detector using dlib.
- But when i give more then 1000 images, data is not trained.Process is killed by system.I have searched online but didn't find any particular solution for that.
- I want to know that is there any particular pattern for giving input images or any other thing that i have to consider for training data.Is there any requirement of RAM or type of images or size of images.
- Any help would be appreciated.Thanks in advance!!!
Asked
Active
Viewed 1,126 times
6
Jayesh
- 4,755
- 9
- 32
- 62
RONAK CHAUHAN
- 115
- 9
-
Without seeing any kind of code it's impossible to say what the problem might be. Maybe you are just creating a lot of objects, but never destroy them? – Some programmer dude Feb 17 '16 at 13:54
-
I am using dlib library code for training.For that i am using `train_object_detector.cpp` from dlib. – RONAK CHAUHAN Feb 18 '16 at 04:41
-
Here you can find source code for training detector http://dlib.net/train_object_detector.cpp.html – RONAK CHAUHAN Feb 18 '16 at 07:35
-
@RONAKCHAUHAN Are you using C++ codes for training? Then try to get exception using `try` & `catch` block . Process is killed by system, it means it is not exiting normally. – Jun 16 '16 at 09:08
-
Sounds like an issue with RAM. See this question: http://stackoverflow.com/questions/39024165/d-lib-object-detector-training/39024492#39024492 – bslawski Oct 12 '16 at 05:09