biomedia-mira / causal-gen Goto Github PK

Hello,

x = transform.inv(y)
event_dim += transform.domain.event_dim - transform.codomain.event_dim
log_prob = log_prob - _sum_rightmost(
transform.log_abs_det_jacobian(x, y),
event_dim - transform.domain.event_dim, )
y = x

Why is x passed as argument to the method log_abs_det_jacobian? It is supposed that this function takes only one argument.

Also, why does this function implement the abs det jacobian of the gumbel transformation? In which part is p_x(g-1(y)) implemented?

Thanks in advance!

Could you release the test code for testing models according to MAE? I urgently need to conduct some experiments with your code. Great thank you for your kind help!

Hello,
Where could I find the version of the ColorMNIST dataset that is produced according to the SCM which is described in this paper. Is ColorMNIST available in this repository?

Thank you in advance!

Hello again,

Where could I find the code that reproduces your results on MorphoMNIST?

Thanks!

How to run experiments on Morpho-MNIST dataset in code

Hi, sir,thanks for the implementation! I'd think it is a great work!

I am little confused with specific counterfactual training steps. It seems that the network in "train_cf.py" only take single counterfactual parent into account during training. But there are actually multi counterfactual parents when infering(Like the case shown in Hugging Face).

So should I run the "train_cf.py" for every counterfactual parent (thickness->intensity->digit) one by one during training ?

Hello,
I wanted to ask why the counterfactual cf_scale is multiplied with a temperature value t_u=0.1 in eval_example.ipynb. Would it be possible to use other t_u values as well? Is there any criterion to determine the best temp value?

t_u=0.1
cf_loc, cf_scale = vae.forward_latents(z, parents=_cf_pa)
cf_scale = cf_scale * t_u

Thanks in advance!