Jonas Dehning comments

Results 14 comments of


                                            Jonas Dehning

Add event_ndims to Transform class to fix shape errors in non-autobatching mode

There is still an error with the notebooks tests. I don't know exactly how to best fix it

Add event_ndims to Transform class to fix shape errors in non-autobatching mode

The notebook creation is fixed in the makefile. Works now but I am not 100\% sure that this is the correct fix.

Error in transformed variables with manual model vectorization

I think I have some ideas on how to fix it. Probably next week I will have some time to look deeper into it and open a PR, if you...

Add stein variational gradient descent as sampling possibility, for increased performance

We are not hell bent on it. My thought was that variational methods could eventually be parallelized, when we will look at the level of Landkreise. But that I find...

Add stein variational gradient descent as sampling possibility, for increased performance

One this issue, no one is actively working on. So if someone want to have a look... Things learned so far: - the pymc3.ASVGD needs very high temperature, around 2...

Windows user strftime bug

What does the `-` in `%-d` do? I would simply remove it, if it is not important...

Merge mobility analysis with master

Good :+1:

Implement `@as_jax_op` to wrap a JAX function for use in PyTensor

I have a question, where should I put the @as_jax_op. Currently, it is in a new file `pytensor/link/jax/ops.py`. Does that make sense? Also, how should one access it? Only by...

Implement `@as_jax_op` to wrap a JAX function for use in PyTensor

> Big level picture. What's going on with the flattening of inputs and why is it needed? To be able to wrap JAX function that accept pytrees as input. `pytensor_diffeqsolve...

Implement `@as_jax_op` to wrap a JAX function for use in PyTensor

> > > Big level picture. What's going on with the flattening of inputs and why is it needed? > > > > > > To be able to wrap...