💳 For enroll in the course you must create your user account first for enter to platform and then
pay through paypal or credit card, once you pay the platform enrolls you automatically in the course. 💳
🎫 Enter to course link, create new account and then pay, if you have any problem contact immediately to:
WhatsApp: 057 + 3167467218
- Complete process about different deep learning object detectors creation and training (mobile and desktop)
- Install, compile and use DLIB Desktop(CPU, GPU)
- Install and use CUDA + CUDNN for GPU training and inference aceleration (Nvidia videocard needed)
- Install and use Visual studio Community for Compile and develop AI programs(DLIB)
- Use imglab for data anotation (image labeling) in DLIB format
- Use imglab for data splitting (train and test)
- Install and use imglabaug (python) fork for DLIB object detection(data augmentation)
- All proccess for train multiclass object detectors with DLIB (image labeling, Parameter tunning, set the xmls path on debug arguments, data augmentation, etc)
- Tuning parameters, arguments and neural network arquitecture elements for speed up the training and inference time (DLIB mobile)
- Define and use especialized custom neural network arquitecture (exclusive own solution) to solve the biggest problem of "tiny object detection" for detect and recognize very small things on images.
- Deserialize and use neural networks trained in object detection program (DLIB .Dat)
- Install and use VirtualBOX to mount Linux ubuntu 18.04 to compile OpenBlas and Android Dlib IA Projects
- Generate standalone android toolchain in linux for compile OpenBlas and generate the .so and .a files to use for Android object detection app
- Compile, install and use OpenBLAS in linux for mobile inference aceleration (acelereration for ARMv7-32 bits, ARMv8-64bits and especial Cortex arquitecture)
- Install and use NDK (Android C, C++ compiler) in Android studio for mobile DLIB object detection apps using big desktop models (not quantized or low precision lite versions for mobile)
- Use DLIB and link with OpenBLAS to acelerate and run your own object detection models in Android studio(Camera multiclass object detection own app)
- Install and use intel MKL + ThreadBB for CPU training/inference aceleration
- All proccess for train multiclass object detectors with TensorFlow for mobile apps.
- Use tensorflow lite trained models in real time on android.
- Install and use Jupyter noteboks for interactive training, eval and testing of Tensorflow models (Python)
- Use google Colab for interactive training, eval and testing of Tensorflow models online if you not have GPU (Python)
- Install and use anaconda for create tensorflow GPU enviroments(Drivers, CUDA, CUPTI, etc) automatically.
- Run and test tensorflow lite trained models in desktop or mobile.
- Convert formats with exclusive own program (DLIB to Pascal VOC and Pascal VOC to DLIB) for interchange plataforms and make diferent projects easy (DLIB and Tensorflow).
Requirements for course:
- Only basic programing experience in any languaje.
- (Optional) Nvidia Graphic Card with at least 4gb of Vram for training proccess with CUDA (GTX, RTX, Quadro, etc)
- Want to learn a lot!
- Access to our own moodle education platform for ordered and easy lessons learning (access for life)
- Very practical lessons for understand and produce your own object detectors for you or your enterprise
- All special architectures for train different type objects (for size, amount of data, form, etc)
- All source code of all programs used in the course lessons
- Step by step lessons and tutorials for easy learning and understanding of topics.
- Very exclusive information and programs not available on the net.
- Knowledge, secrets and experience for deep learning object detectors reflected in the lessons.
- Direct contact with teacher (AI MSc. Sergio Andrés Gutiérrez Rojas)