UCCSD Refactoring #66
Merged
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
This adds some drafts of the classes we plan to add during the refactoring of UCCSD. More details can be found in qiskit-community#57 (comment)
mrossinek
commented
Feb 26, 2021
|
Can we put the circuits in |
All locations and namings are very temporary. I just wanted to get this out there. But yes, this is definitely something we can do! |
woodsp-ibm
reviewed
Feb 26, 2021
woodsp-ibm
reviewed
Feb 26, 2021
qiskit_nature/components/variational_forms/excitation_op_builder.py
Outdated
Show resolved
Hide resolved
woodsp-ibm
reviewed
Feb 26, 2021
This is an early attempt on what I think UCC may look like. I am still unsure where to place the `ExcitationOpBuilder` in the code.
Cryoris
reviewed
Mar 3, 2021
woodsp-ibm
reviewed
Mar 4, 2021
qiskit_nature/circuit/library/ansaetze/evolved_operator_ansatz.py
Outdated
Show resolved
Hide resolved
qiskit_nature/circuit/library/ansaetze/evolved_operator_ansatz.py
Outdated
Show resolved
Hide resolved
qiskit_nature/circuit/library/ansaetze/evolved_operator_ansatz.py
Outdated
Show resolved
Hide resolved
Co-authored-by: Julien Gacon <gaconju@gmail.com> Co-authored-by: Steve Wood <40241007+woodsp-ibm@users.noreply.github.com>
Rather than duplicating the adaptive nature of an Ansatz through the introduction an additional interface we can rely on an adaptive algorithm to handle the list of excitations. Currently (AdaptVQE) does this anyways and the `push` and `pop` methods were merely utilities. Now, the BlueprintCircuit can simply have its operators replaced which will automatically invalidate a cached circuit and will cause rebuilding of all relevant information.
To clarify: I did not mean to outright remove it, merely to postpone it until a feature release in the future. |
We will refactor MP2Info in the future and add it as part of a feature release.
…e into uccsd-refactoring
…e into uccsd-refactoring
woodsp-ibm
reviewed
Mar 22, 2021
mrossinek
commented
Mar 22, 2021
4 tasks
pbark
approved these changes
Mar 25, 2021
|
@woodsp-ibm I know that you did not have time to review this yesterday but we need to continue with the algorithms where we rely on this work. Thus I merged it. |
mrossinek
added a commit
to mrossinek/qiskit-nature
that referenced
this pull request
Mar 25, 2021
* Add drafts of classes for UCCSD refactoring This adds some drafts of the classes we plan to add during the refactoring of UCCSD. More details can be found in qiskit-community#57 (comment) * Fix mypy, spelling, lint * add rough draft of EvolvedOpAnsatz * Move EvolvedOpAnsatz to circuit.library.ansaetze * Add rough draft of UCC This is an early attempt on what I think UCC may look like. I am still unsure where to place the `ExcitationOpBuilder` in the code. * Some improvements based on code reviews Co-authored-by: Julien Gacon <gaconju@gmail.com> Co-authored-by: Steve Wood <40241007+woodsp-ibm@users.noreply.github.com> * Update ExcitationBuilder * Move adaptive Ansätze to circuit.library submodule * Add ideas for unittests * Remove AdaptiveAnsatz interface Rather than duplicating the adaptive nature of an Ansatz through the introduction an additional interface we can rely on an adaptive algorithm to handle the list of excitations. Currently (AdaptVQE) does this anyways and the `push` and `pop` methods were merely utilities. Now, the BlueprintCircuit can simply have its operators replaced which will automatically invalidate a cached circuit and will cause rebuilding of all relevant information. * Fix excitation operators * Assert qubit_op_converter during config checking * Fix unittest for ExcitationBuilder * Replace dummy QubitOpConverter class * Fix linters * Preserve excitation order We must preserve the excitation order to enforce unique operators during separate runs. By using lists rather than sets, this can be easily achieved. * Add arguments to filter excitations by spin type * Move circuit/library tests into submodules * Expose filter keyword arguments in UCC class * Remove uppercase from spelling dict * Fix error message * Add UCC to circuit.library module imports * Add UCCSD convenience subclass of UCC * Minor update to UCC unittests * Fix spell * Rename "ansaetze" -> "ansatzes" * Replace `pure_spin` with `max_spin_excitation` The original intent of `pure_spin` was to add an easy configuration option to specify the SUCC Ansatz. However, the original implementation was backwards: `pure_spin` removed the alpha,beta double excitations but in fact, we want to remove the alpha,alpha and beta,beta ones. In the new implementation we have the optional integer argument called `max_spin_excitation`. This can be used to set the maximum number of excitations within identical spin orbitals! For example, setting `num_excitations=2` will compute all possible double excitations. However, in combination with `max_spin_excitation=1` this list will be filtered as to only include those where per spin, we perform at most 1 excitation. * Move build_excitation_ops outside of a class * Various minor improvements * Do not include permutations of excitations * Replace UCCSD in VQEUCCSDFactory * Fix spell * Improve docstrings * Fix spell * Split excitation operator builder Rather than generating the excitations and building the operators within the same method, we split this into two parts for increased flexibility. * Add the SUCCD convenience subclass * Add the PUCCD convenience subclass * Extract num_particles validation * Fix mypy * Add single excitation inclusion options * Add some debug-level logging * Make excitation genertor fermionic by naming * Fix docstring * WIP: vibrational mode excitation generator * Fix spell * Actually fix spell * Update docstrings * Add/cleanups unittests for excitation generators * Improve UCCSD test * Add proper tests for PUCCD and SUCCD * Add more tests for plain UCC * Add spell exceptions * Make mypy happy * Fix spell * Fix unittest * Fix operator caching * Fix spell * Remove MP2Info We will refactor MP2Info in the future and add it as part of a feature release. * Add initial_state to UCC * Fix after merge * Add preferred_init_points * add tests * Fix lint * evolved op tests * fix spell * Move initial_state property to EvolvedOpAnsatz * Ensure PauliTrotterEvolution is used for UCC * Move preferred_init_points property to EvolvedOpAnsatz Co-authored-by: Manoel Marques <manoel.marques@ibm.com> Co-authored-by: Julien Gacon <gaconju@gmail.com> Co-authored-by: Steve Wood <40241007+woodsp-ibm@users.noreply.github.com>
mrossinek
added a commit
to mrossinek/qiskit-nature
that referenced
this pull request
Mar 25, 2021
* Add drafts of classes for UCCSD refactoring This adds some drafts of the classes we plan to add during the refactoring of UCCSD. More details can be found in qiskit-community#57 (comment) * Fix mypy, spelling, lint * add rough draft of EvolvedOpAnsatz * Move EvolvedOpAnsatz to circuit.library.ansaetze * Add rough draft of UCC This is an early attempt on what I think UCC may look like. I am still unsure where to place the `ExcitationOpBuilder` in the code. * Some improvements based on code reviews Co-authored-by: Julien Gacon <gaconju@gmail.com> Co-authored-by: Steve Wood <40241007+woodsp-ibm@users.noreply.github.com> * Update ExcitationBuilder * Move adaptive Ansätze to circuit.library submodule * Add ideas for unittests * Remove AdaptiveAnsatz interface Rather than duplicating the adaptive nature of an Ansatz through the introduction an additional interface we can rely on an adaptive algorithm to handle the list of excitations. Currently (AdaptVQE) does this anyways and the `push` and `pop` methods were merely utilities. Now, the BlueprintCircuit can simply have its operators replaced which will automatically invalidate a cached circuit and will cause rebuilding of all relevant information. * Fix excitation operators * Assert qubit_op_converter during config checking * Fix unittest for ExcitationBuilder * Replace dummy QubitOpConverter class * Fix linters * Preserve excitation order We must preserve the excitation order to enforce unique operators during separate runs. By using lists rather than sets, this can be easily achieved. * Add arguments to filter excitations by spin type * Move circuit/library tests into submodules * Expose filter keyword arguments in UCC class * Remove uppercase from spelling dict * Fix error message * Add UCC to circuit.library module imports * Add UCCSD convenience subclass of UCC * Minor update to UCC unittests * Fix spell * Rename "ansaetze" -> "ansatzes" * Replace `pure_spin` with `max_spin_excitation` The original intent of `pure_spin` was to add an easy configuration option to specify the SUCC Ansatz. However, the original implementation was backwards: `pure_spin` removed the alpha,beta double excitations but in fact, we want to remove the alpha,alpha and beta,beta ones. In the new implementation we have the optional integer argument called `max_spin_excitation`. This can be used to set the maximum number of excitations within identical spin orbitals! For example, setting `num_excitations=2` will compute all possible double excitations. However, in combination with `max_spin_excitation=1` this list will be filtered as to only include those where per spin, we perform at most 1 excitation. * Move build_excitation_ops outside of a class * Various minor improvements * Do not include permutations of excitations * Replace UCCSD in VQEUCCSDFactory * Fix spell * Improve docstrings * Fix spell * Split excitation operator builder Rather than generating the excitations and building the operators within the same method, we split this into two parts for increased flexibility. * Add the SUCCD convenience subclass * Add the PUCCD convenience subclass * Extract num_particles validation * Fix mypy * Add single excitation inclusion options * Add some debug-level logging * Make excitation genertor fermionic by naming * Fix docstring * WIP: vibrational mode excitation generator * Fix spell * Actually fix spell * Update docstrings * Add/cleanups unittests for excitation generators * Improve UCCSD test * Add proper tests for PUCCD and SUCCD * Add more tests for plain UCC * Add spell exceptions * Make mypy happy * Fix spell * Fix unittest * Fix operator caching * Fix spell * Remove MP2Info We will refactor MP2Info in the future and add it as part of a feature release. * Add initial_state to UCC * Fix after merge * Add preferred_init_points * add tests * Fix lint * evolved op tests * fix spell * Move initial_state property to EvolvedOpAnsatz * Ensure PauliTrotterEvolution is used for UCC * Move preferred_init_points property to EvolvedOpAnsatz Co-authored-by: Manoel Marques <manoel.marques@ibm.com> Co-authored-by: Julien Gacon <gaconju@gmail.com> Co-authored-by: Steve Wood <40241007+woodsp-ibm@users.noreply.github.com>
Anthony-Gandon
pushed a commit
to Anthony-Gandon/qiskit-nature
that referenced
this pull request
May 25, 2023
* Add drafts of classes for UCCSD refactoring This adds some drafts of the classes we plan to add during the refactoring of UCCSD. More details can be found in qiskit-community#57 (comment) * Fix mypy, spelling, lint * add rough draft of EvolvedOpAnsatz * Move EvolvedOpAnsatz to circuit.library.ansaetze * Add rough draft of UCC This is an early attempt on what I think UCC may look like. I am still unsure where to place the `ExcitationOpBuilder` in the code. * Some improvements based on code reviews Co-authored-by: Julien Gacon <gaconju@gmail.com> Co-authored-by: Steve Wood <40241007+woodsp-ibm@users.noreply.github.com> * Update ExcitationBuilder * Move adaptive Ansätze to circuit.library submodule * Add ideas for unittests * Remove AdaptiveAnsatz interface Rather than duplicating the adaptive nature of an Ansatz through the introduction an additional interface we can rely on an adaptive algorithm to handle the list of excitations. Currently (AdaptVQE) does this anyways and the `push` and `pop` methods were merely utilities. Now, the BlueprintCircuit can simply have its operators replaced which will automatically invalidate a cached circuit and will cause rebuilding of all relevant information. * Fix excitation operators * Assert qubit_op_converter during config checking * Fix unittest for ExcitationBuilder * Replace dummy QubitOpConverter class * Fix linters * Preserve excitation order We must preserve the excitation order to enforce unique operators during separate runs. By using lists rather than sets, this can be easily achieved. * Add arguments to filter excitations by spin type * Move circuit/library tests into submodules * Expose filter keyword arguments in UCC class * Remove uppercase from spelling dict * Fix error message * Add UCC to circuit.library module imports * Add UCCSD convenience subclass of UCC * Minor update to UCC unittests * Fix spell * Rename "ansaetze" -> "ansatzes" * Replace `pure_spin` with `max_spin_excitation` The original intent of `pure_spin` was to add an easy configuration option to specify the SUCC Ansatz. However, the original implementation was backwards: `pure_spin` removed the alpha,beta double excitations but in fact, we want to remove the alpha,alpha and beta,beta ones. In the new implementation we have the optional integer argument called `max_spin_excitation`. This can be used to set the maximum number of excitations within identical spin orbitals! For example, setting `num_excitations=2` will compute all possible double excitations. However, in combination with `max_spin_excitation=1` this list will be filtered as to only include those where per spin, we perform at most 1 excitation. * Move build_excitation_ops outside of a class * Various minor improvements * Do not include permutations of excitations * Replace UCCSD in VQEUCCSDFactory * Fix spell * Improve docstrings * Fix spell * Split excitation operator builder Rather than generating the excitations and building the operators within the same method, we split this into two parts for increased flexibility. * Add the SUCCD convenience subclass * Add the PUCCD convenience subclass * Extract num_particles validation * Fix mypy * Add single excitation inclusion options * Add some debug-level logging * Make excitation genertor fermionic by naming * Fix docstring * WIP: vibrational mode excitation generator * Fix spell * Actually fix spell * Update docstrings * Add/cleanups unittests for excitation generators * Improve UCCSD test * Add proper tests for PUCCD and SUCCD * Add more tests for plain UCC * Add spell exceptions * Make mypy happy * Fix spell * Fix unittest * Fix operator caching * Fix spell * Remove MP2Info We will refactor MP2Info in the future and add it as part of a feature release. * Add initial_state to UCC * Fix after merge * Add preferred_init_points * add tests * Fix lint * evolved op tests * fix spell * Move initial_state property to EvolvedOpAnsatz * Ensure PauliTrotterEvolution is used for UCC * Move preferred_init_points property to EvolvedOpAnsatz Co-authored-by: Manoel Marques <manoel.marques@ibm.com> Co-authored-by: Julien Gacon <gaconju@gmail.com> Co-authored-by: Steve Wood <40241007+woodsp-ibm@users.noreply.github.com>
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.
Summary
Closes #57, closes #51, closes #55
Details and comments
For more details see the discussion there. I will update this description as we go.