Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Update L_op of Blockwise #3

Closed
wants to merge 38 commits into from
Closed

Update L_op of Blockwise #3

wants to merge 38 commits into from

Conversation

purna135
Copy link

This PR adds:

  • get_output_info(), which is the same as Elemwise get_output_info(), to make all inputs of the same dimension.
  • derive DimShuffle's gufunc signature
  • reduce the broadcasted dimensions of inputs after the grad is computed

Things changed in this PR:

  • instead of using the dimensions from the start for computing the curr_static_shape of core_inp_grads use the dimensions from the end.
  • an extra check before calling perform() of DimShuffle (which can be removed later)

rlouf and others added 30 commits September 2, 2022 10:03
@rlouf @brandonwillard
Make ChoiceRV behave like NumPy's choice
78706bd 
`ChoiceRV` currently does not behave like its NumPy equivalent when its
`a` parameters is an `int` or an array with more than one dimension. We
implement the necessary changes so that it does.
@rlouf @brandonwillard
Remove unnecessary cloning of NoneConst variable
c2ed818
@ltoniazzi
Add tensor conversion to flatnonzero, nonzero_values, tile, `in…
e40c827 
…verse_permutation`, and `diag`
@ricardoV94 @brandonwillard
Fix bug in static type shape of DiagonalSubtensor Op output
70cf7e3
@ricardoV94 @brandonwillard
Use static shapes in outputs of Elemwise
ddd7322
@ricardoV94 @brandonwillard
Allow type of shared variable updates to have more specialized shape …
1d899ee 
…information

Otherwise TestSharedOptions.test_specify_shape_inplace would fail
@rlouf @brandonwillard
Refactor the JAX dispatcher
74db799
@rlouf @brandonwillard
Fix typing of RewriteDatabaseQuery arguments
6668f44
@rlouf @brandonwillard
Ignore attr-define mypy error
7075c7f
@eganster
Update options.package_data section of setup.cfg
b35d618
@rlouf @brandonwillard
Warn instead of raise when converting CheckAndRaise
1e63a48
@rlouf @brandonwillard
Remove omnistaging warning
0b9bba1
@brandonwillard
Update CheckAndRaise to handle ScalarType inputs
5043ac8
@brandonwillard
Remove SciPy check from aesara.sparse
8528590
@brandonwillard
Fix Type conversion and checking in IfElse.make_node
a5f8b69
@brandonwillard
Allow conversion of dense to sparse types in SparseTensorType.convert…
d3ff35d 
…_variable
@brandonwillard
Clean up comments and docstrings in aesara.sparse
373eab1
@brandonwillard
Refactor SparseTensorType constructor and clone interface
3e5ad99
@brandonwillard
Make SparseTensorType.filter match TensorType in error behavior
7d0edb8
@ricardoV94 @brandonwillard
Allow TensorType convert_variable to work with mixed static shapes
84e69fc
@ricardoV94 @brandonwillard
Fix type inference in Elemwise when inputs have 0 shape
e537311
@ricardoV94 @brandonwillard
Check Split axis is a scalar
72e67d8
@ricardoV94 @brandonwillard
Fix Split output type
cd3a3ce
@rlouf @brandonwillard
Use unique input names in numba_funcify_Scan
0d69809
@ricardoV94
Implement moveaxis
5d97ffa
@rlouf
Add deprecation warning for rng_mrg
864ee33
@ricardoV94 @brandonwillard
Add tests for shared updates in JAX and NUMBA backends
3aff3d2 
* Also fixes bug in JITCompiler when first output of inner fgraph is an input variable, as can happen in some specific functions with updates
@brandonwillard
Update pre-commit hooks
b73abda
@brandonwillard
Make isort ignore Cython files
8d61202
@LegrandNico
Update Scan Gibbs example in documentation
9a85dbc
@brandonwillard
Copy link
Owner

Closed in favor of aesara-devs#1215.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
8 participants
@purna135 @brandonwillard @rlouf @ltoniazzi @ricardoV94 @eganster @LegrandNico @kc611