GSOC community bonding week 2: The Paper...s
"Previously on LibreHealth GSOC 2020" [ Recap ]
"Django" [ Gitlab ]
The steps to setup your own rest authentication service using Django:
Install django rest auth and allauth:
pip install django-allauth
Signup with a smtp service provider and get the credentials.
Modify your settings.py
First the installed apps:
INSTALLED_APPS = [ ... 'rest_framework.authtoken', 'rest_auth', 'django.contrib.sites', 'allauth', 'allauth.account', 'rest_auth.registration', 'allauth.socialaccount' ... ]
SMTP server keys:
EMAIL_HOST = 'smtp-relay.sendinblue.com' EMAIL_PORT = 587 EMAIL_USE_TLS = True EMAIL_USE_SSL = False EMAIL_HOST_USER = 'singh.kislay.kunal@gmail.com' EMAIL_HOST_PASSWORD = os.environ.get('EMAIL_HOST_PASSWORD')
Modify your urls.py
url_pattern = [ ... re_path(r'^rest-auth/registration/account-email-verification-sent/', views.null_view, name='account_email_verification_sent'), re_path(r'^rest-auth/registration/account-confirm-email/(?P<key>[-:\w]+)/$', ConfirmEmailView.as_view(), name='account_confirm_email'), re_path(r'^rest-auth/registration/complete/$', views.complete_view, name='account_confirm_complete'), re_path(r'^password-reset/confirm/(?P<uidb64>[0-9A-Za-z_\-]+)/(?P<token>[0-9A-Za-z]{1,13}-[0-9A-Za-z]{1,20})/$', views.null_view, name='password_reset_confirm'), re_path(r'rest-auth/', include('rest_auth.urls')), re_path(r'^rest-auth/registration/', include('rest_auth.registration.urls')) ... ]
Easy, you have your authentication server setup. Don't forget to make the migrations.
"Paper...s" [ Reading ]
There were several papers that were mentioned to be read and one of them a had a public code repo, so can you guess which paper I started with? I started with "Deep Neural Networks Improve Radiologists’Performance in Breast Cancer Screening" This is simply a beautiful paper, the idea of getting heat maps by sliding window over 224*224 chunks of image and then sending these heat maps directly in to a different neural nework that accepts 4 images, each image being the image(grayscale), heatmap of the malign and heatmap of the benign area for making 3 channels is amazing. mind == blown. I loved reading the paper, right now I'm working on porting there code to tensorflow 2.1 as it is in tensorflow 1.13. Why you ask? well, new is always better.
"Next week" [ Plan ]
Port the model to tf 2.1. I have never ported a tf script to a different versions, this will be new and fun. Right now I just feel like I can do it.
Comments
Post a Comment