tranphuminhbkhn / MiLu

Machine Learning /Deep Learning framework, implemention in C++ using MPI
1 stars 0 forks source link

question #1

Open dimiy opened 4 years ago

dimiy commented 4 years ago

sorry can you give your email address or where is it convenient for you to chat? I wanted to ask something about the framework and skills.

tranphuminhbkhn commented 4 years ago

Hi, my email is: tranphuminhkhn@gmail.com

dimiy commented 4 years ago

I get a message that your address was not found so I will ask a question here I hope that you do not mind can you tell me resources where it is shown writing procts in deep learning from scratch (without frameworks) and where neural entities write from scratch? Express your opinion on the given resources. you are welcome. http://blog.ezyang.com/2019/05/pytorch-internals/ https://medium.com/@whitesell.ben/how-to-build-your-own-deep-learning-framework-its-easy-i-ll-prove-it-c859cb790386 https://end-to-end-machine-learning.teachable.com/p/write-a-neural-network-framework/ https://www.xinjianl.com/blog/tag/pytensor/page/4/ http://dlsys.cs.washington.edu/ https://course.fast.ai/videos/?lesson=8 is it possible to implement object recognition from scratch (with its own framework)? If so, what kind of project architecture do you need? many thanks if you can help

tranphuminhbkhn commented 4 years ago

Hi, Sorry, I mistakenly wrote. My email is: tranphuminhbkhn@gmail.com.

And about your question. I don't write neural from scratch, but tensor. You can see src/tensor.cpp, where i define tensor class. Because it's a course projects, a simple way i just define 2d tensor.

Tensor class include: Properties

Methods

And of course, it is possible to implement object recognition from scratch. But my framwork not, I don't write convolutional layer, because it's quite complicated.

Let me know if you have any questions.