Pyro Discussion Forum
Aug 14, 2025 · Forum For Pyro Developers
Bad performance Bayesian Convolutional Neural Network - Misc.
Apr 28, 2025 · Hello Pyro community, I’m trying to build a Bayesian CNN for MNIST classification using Pyro, but despite seeing the ELBO loss decrease to around 10 during training, the …
numpyro - Pyro Discussion Forum
Jun 3, 2019 · Forum For Pyro Developers
Extending HMM example with additional independent dimension
Feb 13, 2019 · HI, I tried to extend the HMM example (at the end on the tutorial on enumeration) by adding additional independent dimensions (e.g data now has a shape num_steps x …
NAN log_prob_sum in Poisson model? - Pyro Discussion Forum
Jun 10, 2018 · Hello Pyro community! I’m a Pyro noob, just getting my feet wet with some simple models, and ran into a warning that I could use some help understanding. I’m attempting to …
Trouble Initializing Parameters for SVI - forum.pyro.ai
Apr 22, 2025 · the issue was because this was not differentiable! the fix was following guidance from Best practice to handle division by zero in auto-differentiation · Issue #5039 · jax-ml/jax · …
GPU available but slow compared to CPU on BNN example
Nov 25, 2021 · Well, It is a VI + NeuraReparametrisation to perform a NUTS. Currently, we have found that the job runs 100% on CPU while GPU memory is activated. it’s as if the CPU was …
Save MCMC results - Misc. - Pyro Discussion Forum
Sep 30, 2018 · Hi - I’m contributing to a library for bayesian analysis and visualization (arviz), and we are trying to add a function to convert Pyro objects into data that can be used there. As …
Implementation & normalizing flow in matrix normal distribution
Nov 1, 2024 · Hi, I’m working on a model where the likelihood follows a matrix normal distribution, X ~ MN_{n,p} (M, U, V). I’m using conjugate priors: M ~ MN U ~ Inverse Wishart V ~ Inverse …
Reducing MCMC memory usage - numpyro - Pyro Discussion Forum
Oct 15, 2023 · I am running NUTS/MCMC (on multiple CPU cores) for a quite large dataset (400k samples) for 4 chains x 2000 steps. mcmc.run actually ran until the end, but then died with an …