Kayıtlar

Week 5 : Stanford cs231 and KNN

So Ive started with this Convnets course on Stanford University. I watched the first three videos. Turned out to be just the thing that I am looking for. They also have some homeworks and all, so I will be able to test myself and put some milestones on the way. Best part of it is the fact that one of the objectives that we have to accomplish as a part of this course is gaining the ability code our own convolutional neural networks from scratch. Thats encouraging. To start with the first assignment, I downloaded Anaconda. It is a tool to manage your different python installations. It makes it possible to manage them in different environments which you can activate and deactivate. I shouldda discover it a long time ago. In assignments, Stanford provides some requirement.txt s that give us the exact state of the environment that our code needs to run. to create an environment and to use it is fairly easy. you just have to go conda create -- name myenv   and   source activ

Week 2 : Reproducing the Results of Deeply Learned Attributes for Crowd Scene Understanding (Jing Shao et al)

Resim
Getting started with www_crowd 1- Downloaded all the data related to Jing Shao's aforementioned research . 2- Cloned the repository in GitHub. 3- Watched the video regarding the research. 4- Downloaded the .caffe models and prototype.txt given. Learning how to train my first network 5- As a tutorial, I read the official siamese network tutorial of caffe. terms that I should be familiar with (I will update what they mean once I learn):      1) Convolution The first layer in a CNN is always a Convolutional Layer . [1] In signal processing, to convolve means to create a function using two others. [2] In this case, convolution refers to the process of multiplication of the image with the kernel in a sliding manner . The kernel will be multiplicated with a portion of the image, which we call as the receptive field , and then will slide by one pixel to do it again. Output of the each multiplication will be only one integer. But this is not a standard matrix-multiplic

Week 1 : Installing Caffe on GTX 1050 Ubuntu 16.04

Install the Graphics - Installed the graphics card on the motherboard. - Bought a converter for card - monitor compatibility (50 rm) - Installed Ubuntu 16.04 LTS 64 bit OS for stability. (It is being said that 17.04 still has issues with CUDA) - From OS settings -> Software & Updates -> Additional Drivers -> NVIDIA Corporation -> Using NVIDIA binary driver   |   |---> To cope with the wrong driver problem   |   |---> Ubuntu 16.04 using Gallium 0.4 on llvmpipe (LLVM 3.8, 256 bits) was being shown before I changed it. - Installed MATLAB Install the Dependencies - Installed caffe dependencies : glog protobuf leveldb snappy opencv hdf5-serial boost blas cuda 9.0 gflags lmdb While installing cuda, script prompted to install nvidia 384 driver which I posess a newer version of. I purged my own driver, installed the one it needs. And only then I was able to install CUDA 9.0 Under BLVC Wiki I followed Ubuntu 16.04 Installation Guide to so