Paarth Dudani

What you suggested does seem to help but it doesn't fully address the issue: 1. SNPE 2. SNRE 3. SNLE I changed the no. of simulations to 1500/2000

Thanks for your suggestions! I have made the appropriate changes. In particular, I rewrote the simulator function, since it was not giving me the noisy data that I was looking...

In this case, should I reduce the dimensionality of the dataset? The reason I am asking this is because according to the following paragraph from the sbi package website, my...

Hi Michael! I am pleasantly surprised by your prompt response! I did make the first change as you had suggested and it did resolve the issue! However, I am now...

The plot is for the posterior on the scale parameter for the exponential being learned by SNLE/SNRE as shown in the code below for SNLE, against the true value of...

Thank you so much for your suggestions! I did try them out. However, the estimates that I am getting from each of the 3 algorithms (SNPE_C, SNLE, SNRE_B) are somewhat...

I am sorry for not attaching the code for this issue, so here goes: 1. SNRE: ``` exp_prior_mod = WrappedExponential(torch.tensor([2.0])) simulator = lambda theta: ((1/theta)*np.exp(-t/theta) + 0.05*torch.rand_like(theta)).float() x_o = torch.tensor(noisy_exponential(theta_true,t)).float()...