-
Notifications
You must be signed in to change notification settings - Fork 29
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.
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
Devel #2578
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Merge pull request from PrincetonUniversity/master
Add missing pytest.mark.composition to tests that construct and run Composition. Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
Pulls in pytorch[0] so it needs the same constraints. [0] https://github.com/ModECI/MDF/blob/main/setup.cfg Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
… bumped package before installing PNL (#2481) Some packages may pull in updated dependencies that cause issues when rolled back. Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
…p packages (#2482) The packages should be rebuilt if some of the shared dependencies change Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
…for not installed packages (#2483) This happens if the package is an 'extras' dependency. e.g. a 'doc' or 'cuda' dependency not installed in 'dev' CI jobs. Fixes package bumps of docs dependencies. Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
Many psyneulink Functions must be split into several MDF functions/expressions to reproduce their behavior. Instead of adding these supplemental functions on the owners of Functions, add Function._assign_to_mdf_model that adds anything necessary to the owner
MDF: support DriftDiffusionIntegrator
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
…on test Pass the modulator value directly from input. Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
…llvm_is_finished_cond' Passing base parameters matches invocation by the scheduler. Any modulation needs to be applied explicitly. Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
…threshold. Matches Python semantics. Add threshold modulation test for integrator based DDM. Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
…ant (#2490) Consistently use modulated mechanism parameters (e.g. threshold) of TransferMechanism in is_finished function whether it's called from inside the mechanism loop, or by the scheduler. Use modulated function parameters in DDM's 'is_finished' function in both internal and scheduler invocation.
scipy doesn't provide 32-bit wheels for >=1.9.2 Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
…0,<1.13.0 (#2504) Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
…2505) The only supported evaluation type atm is "evaluate_type_objective", which uses the result of ocm.objective_mechanism as output (combined with costs). Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
jupyter pulls in jupyter-server-terminals which requires pywinpty>=2.0.3 on windows [0] [0] https://github.com/jupyter-server/jupyter_server_terminals/blob/main/pyproject.toml Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
…tation (#2560) Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
…anch name (#2566) Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
…d in MechanismError (#2569) string values of errors are not stable and can change between versions. Fixes 'input_list_of_strings' tests when using numpy>=1.22.x. Refactor handling of input port execution exceptions to only check calls to 'execute', parameter set/get do not throw exceptions. Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
…tions Individual terms of the sum are close in magnitude. Instead of accmulating all the postive terms, alternate between positive and negative terms. This keeps the intermadiate/partial results small in magnitude improving accuracy of the floating point expression. This is enough to make the "test_drift_difussion_analytical_shenhav_compat_mode" when using numpy-1.22 with its new AVX512 based floating point operations [0] [0] numpy/numpy@1eff1c5 Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
… operations Individual terms of the sum are close in magnitude. Instead of accmulating all the postive terms, alternate between positive and negative terms. This keeps the intermadiate/partial results small in magnitude improving accuracy of the floating point expression. This matches the updated Python version. Fix use of incorrect terms. Skew calculation needs cubic terms. Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
…22 and AVX512 CPUs Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
Updates the requirements on [numpy](https://github.com/numpy/numpy) to permit the latest 1.22.x release. - [Release notes](https://github.com/numpy/numpy/releases) - [Changelog](https://github.com/numpy/numpy/blob/main/doc/RELEASE_WALKTHROUGH.rst) - [Commits](numpy/numpy@v1.17.0...v1.22.4) Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
…function and update numpy to <1.22.5 (#2476) numpy-1.22 introduced a new set of optimized floating point routines on CPUs that support AVX512 instructions[0]. Reassociate operations in the DriftDiffusionAnalytical function to improve accuracy, this is enough to pass the "test_drift_difussion_analytical_shenhav_compat_mode" test that otherwise fails when using numpy 1.22.x on CPUs that support AVX512 instructions. Update results in tests/functions/test_distribution.py::test_execute[DriftDiffusionAnalytical..] to match the new reassocaited routines for small drift rate and fp32 precision. Use windows/mac expected results on CPUs that support AVX512 when numpy>=1.22 is used. Allow numpy < 1.22.5 Closes: #2572 [0] numpy/numpy@1eff1c5
* • function.py - Function_Base: move arguments referring to parameters in call to function to params arg (to be treated as runtime_params) • memoryfunctions.py - ContentAddressableMemory: add store and retrieve convenience methods * -
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
CodeQL found more than 10 potential problems in the proposed changes. Check the Files changed tab for more details.
* master: Add CodeQL workflow for GitHub code scanning (#2533)
Codeql conflict
This PR causes the following changes to the html docs (ubuntu-latest-3.7-x64):
See CI logs for the full diff. |
jdcpni
approved these changes
Jan 14, 2023
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.