-
Notifications
You must be signed in to change notification settings - Fork 6
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Full batch sensitivity analysis #12
Conversation
@jaklevab |
También está hecha la actualización de los |
Mini-batch sampling
sensitivity_analysis.py
Outdated
dim_ξ_grid=np.linspace(start=1,stop=l,num=3,dtype=np.int64) | ||
n_ξ_samples_grid=np.array([1,20]) | ||
labels = np.zeros((l * k, l), dtype=np.float32) | ||
features = labels + np.abs(np.random.normal(loc=0.0, scale=1.5, size=(l * k, l))).astype(np.float32) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Una preguntita acá: es a propósito que los labels son np.zeros()
, y que entonces los features no tiene correlación con la red?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Emm si no mal recuerdo si. Dijimos que en la primera parte, dejábamos al autoencoder predecir usando información de red y basta. En la segunda parte (que todavía falta) le introducíamos todo el tema de los labels ortogonales a las comunidades a partir de las cuales generamos la red (i.e generar una red de mismas dimensiones y usar las comunidades de la segunda como labels de la primera).
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Okidoki - gracias!
Updated results using the minibatch sampling. The batch size is not tested here and it's influence remains to be seen for the reconstruction loss. The results are similar to the ones obtained when using the full-batch optimization (as the size of the batch is set to consider everything). TODO: Consider smaller batch sizes.
No description provided.