pytensor
pytensor copied to clipboard
pymc-devs
PyTensor allows you to define, optimize, and efficiently evaluate mathematical expressions involving multi-dimensional arrays.
### Description Functions like `scipy.special.psi` now always upcast to float64, even if the input is a low precision integer like `int8`. We need to handle these types, but: 1) float64...
### Description Brought up in #846 ```python import pytensor import pytensor.tensor as pt x = pt.vector("x", dtype="int64") out = pt.special.softmax(x) # Doesn't seem right out.dprint(print_type=True) # Softmax{axis=None} [id A] #...
### Description There are two issues with the code generated by this snippet: ```python def update(x): return pt.exp(x) - 5 x_init = pt.vector("x_init", shape=(7,)) x_init_tangent = pt.vector("x_init_tangent", shape=(7,)) seq, updates...
## Description Try to include most of the functions from slinalg and nlinalg in the documenation ## Related Issue - [ ] Closes #854 ## Checklist - [ ] Checked...
### Description It's rather hard to find what functions are available in PyTensor (like do we have anything for linalg?). We should get sphinx to auto-populate most things with the...
### Description Verify and update support for Numba Ops with the latest versions of Numba. There are some numba issues mentioned in comments that are now closed (as shown below)....
https://github.com/pymc-devs/pytensor/blob/d3bd1f15a497c05a979a8e3e8be40883f669a0b6/pytensor/link/jax/dispatch/elemwise.py#L72-L89 The JAX docs of lax.reshape (which np.reshape uses) suggest this may be better for further optimizations: https://jax.readthedocs.io/en/latest/_autosummary/jax.lax.reshape.html#jax.lax.reshape Relevant part: > For inserting/removing dimensions of size 1, prefer using lax.squeeze...
### Description With static types it is possible to find out cases where we know the reshape is useless at compile time. We should remove it when that's the case....
### Describe the issue: The ```Max``` op which is a subclass of ```CAReduce``` fails for 64 bit unsigned integers. This is also evident in the [PR 731](https://github.com/pymc-devs/pytensor/pull/731#discussion_r1582687763) and its [test](https://github.com/pymc-devs/pytensor/actions/runs/9067607709/job/24913191657?pr=731#step:6:9016)....
### Description Or have rewrites that handle this case. Otherwise it's hard / impossible to use Sparse variables in Blockwise/RandomVariables that have other batched inputs, but would otherwise work fine...