Add dynamic_one_shot tensorflow support + expanded testing
#5861
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
1 task
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## v0.37.0-rc0 #5861 +/- ##
==============================================
Coverage ? 99.66%
==============================================
Files ? 425
Lines ? 40465
Branches ? 0
==============================================
Hits ? 40331
Misses ? 134
Partials ? 0 ☔ View full report in Codecov by Sentry. |
mudit2812
commented
Jun 19, 2024
| @pytest.mark.parametrize("mcm_method", ["one-shot", "tree-traversal"]) | ||
| @pytest.mark.parametrize("shots", [100, [100, 101], [100, 100, 101]]) | ||
| @pytest.mark.parametrize("postselect", [None, 0, 1]) | ||
| def test_sample_with_prng_key(self, mcm_method, shots, postselect): |
There was a problem hiding this comment.
This test and the one below it have just been moved into the TestJaxIntegration class so no need to be thorough reviewing the aforementioned two tests
mudit2812
changed the title
dynamic_one_shot differentiation testsdynamic_one_shot tensorflow support + expanded testing
albi3ro
reviewed
Jun 21, 2024
| mcm_samples = qml.math.array(mcm_samples, like=interface) | ||
| # Can't use boolean dtype array with tf, hence why conditionally setting items to 0 or 1 | ||
| has_postselect = qml.math.array( | ||
| [[int(op.postselect is not None) for op in all_mcms]], like=interface | ||
| [[int(op.postselect is not None) for op in all_mcms]], | ||
| like=interface, |
There was a problem hiding this comment.
I don't think we want to convert back to interface data here. If we call dynamic_one_shot after we've converted to numpy, we don't want to convert back into interface data.
Right now it looks like we are still calling convert_to_numpy_parameters after device preprocessing, but I could very easily see that changing in the future.
I was under the assumption that interface was passed to this function for the purpose of determining jit-specific postselection handling.
There was a problem hiding this comment.
Maybe we could have a "postselection interface" and "result data interface"?
There was a problem hiding this comment.
I was under the assumption that interface was passed to this function for the purpose of determining jit-specific postselection handling.
Yes, for pennylane that is the case, so I can definitely spend time investigating if we really need to convert to interface data, but for Catalyst we need to propagate jax data correctly.
albi3ro
reviewed
Jun 21, 2024
| dev = get_device(shots=shots, seed=PRNGKey(678)) | ||
| params = [np.pi / 4, np.pi / 3] | ||
| obs = qml.PauliZ(0) @ qml.PauliZ(1) | ||
| def test_finite_diff_in_transform_program(interface, mcm_method): |
There was a problem hiding this comment.
I'm not entirely sure this is a mcm test if the diff method is explicitly specified as "finite-diff". What happens if diff_method="best"? Should we automatically switch to finite diff, or give a warning?
There was a problem hiding this comment.
Then we use parameter shift, but parameter shift is not correct if there is postselection or classical control.
There was a problem hiding this comment.
Should we automatically switch to finite diff, or give a warning?
I think it makes sense to switch to finite diff.
**Context:** Currently, `draw()` and `draw_mpl()` can only be requested after applying the full transform program, with the exception of the stages provided through expansion_strategy. **Description of the Change:** Using the new `level` argument from construct_batch, the functions are adapted to make use of the argument as well, which allows for more flexible requests and ability to pinpoint where exactly in the transform program to draw a circuit. As done before, the new functionality works with transforms that split the tape (only in the case for `draw`. For `draw_mpl`, a warning is raised and only the first tape is plotted). **Benefits:** Better plotting UX. **Note for Reviewers:** Minor bugs in `construct_batch` have been discovered during work on this PR, and so expect minor fixes to tests relating to `specs`. [[sc-53735](https://app.shortcut.com/xanaduai/story/53735)] Supersedes #5139 --------- Co-authored-by: Mudit Pandey <mudit.pandey@xanadu.ai> Co-authored-by: Thomas R. Bromley <49409390+trbromley@users.noreply.github.com>
**Context:** Adds a transform for adding `NoiseModels`. **Description of the Change:** Adds `add_noise.py` under `pennylane/transforms` that gives the said method for inserting operations according to a provided noise model. **Benefits:** We support noise models. **Possible Drawbacks:** None. **Related GitHub Issues:** [sc-64843] --------- Co-authored-by: Jay Soni <jbsoni@uwaterloo.ca>
Automatic update of stable requirement files to snapshot valid python environments. Because bots are not able to trigger CI on their own, please do so by pushing an empty commit to this branch using the following command: ``` git commit --allow-empty -m 'trigger ci' ``` Alternatively, wait for this branch to be out-of-date with master, then just use the "Update branch" button! Note that it is expected that the PennyLane-Lightning repo is a version ahead of the release, because the version number is taken from the dev branch. Trying to `pip install` from the files will fail until that major version of Lightning is released. If pip install fails with a not found error when installing because of this, it can be fixed by manually downgrading the PennyLane-Lightning version number in the file by 1 version and trying again. --------- Co-authored-by: GitHub Actions Bot <> Co-authored-by: Christina Lee <christina@xanadu.ai>
**Context:** Adding documentation for the new debugger feature. --------- Co-authored-by: Mikhail Andrenkov <mikhail@xanadu.ai> Co-authored-by: Utkarsh <utkarshazad98@gmail.com> Co-authored-by: Diego <67476785+DSGuala@users.noreply.github.com> Co-authored-by: Josh Izaac <josh146@gmail.com> Co-authored-by: soranjh <40344468+soranjh@users.noreply.github.com> Co-authored-by: Diego <diego_guala@hotmail.com>
…asets (#5484) **Context:** Users have trouble finding all `bondlength` and `basis` combinations for qchem molecules. They end up using the slider on the website to check which bond lengths are available one-by-one.  **Description of the Change:** - Updated the quickstart guide to say "viewing available datasets" instead of "dataset structure" - Updated the `qml.data` docs to encourage users to first read the quickstart guide **Benefits:** Users should be more likely to reach the `list_datasets` description in the quickstart guide and know that it can be used to view the available datasets. **Possible Drawbacks:** Users who use pennylane.ai/datasets but don't go to the documentation still won't see this. --------- Co-authored-by: Utkarsh <utkarshazad98@gmail.com>
**Context:** The PennyLane documentation has a quickstart page on measurements, which includes a large section about mid-circuit measurements. With the focus on dynamic circuits and mid-circuit measurements more generally, we would like to make the content on MCMs more visible. **Description of the Change:** Extracts the MCM part of the measurements quickstart and adds it as its own quickstart page. Adds some details on simulation techniques and polishes the content. **Benefits:** Visibility, documentation extent for MCMs **Possible Drawbacks:** **Related GitHub Issues:** [sc-65318] --------- Co-authored-by: Vincent Michaud-Rioux <vincentm@nanoacademic.com>
**Context:** Previously, the QASM circuit converted to PL had no measurements (even if the circuit ended with measurements, the assumption was we remove those and add PL measurements once the circuit is converted). In the UnitaryHack, a contributor added the option to specify measurements when converting a circuit `from_qasm`, like the `from_qiskit` function. In doing so, the `from_qasm` circuit took on the current `from_qiskit` convention. Calling `from_qasm(qasm_str)` previously returned only the operators, and now returns operators + measurements. This is a breaking change with no deprecation cycle. PennyLaneAI/pennylane-qiskit#469 **Description of the Change:** Adds a deprecation warning to `from_qasm` [sc-66318] --------- Co-authored-by: lillian542 <38584660+lillian542@users.noreply.github.com> Co-authored-by: Christina Lee <christina@xanadu.ai> Co-authored-by: David Wierichs <david.wierichs@xanadu.ai>
**Context:** The `from_openfermion` function is updated to supports the conversion of qubit operators in addition to fermionic operators. **Description of the Change:** **Benefits:** **Possible Drawbacks:** **Related GitHub Issues:**
**Context:** PyTrees are an ideal format for dataset attributes, since they provide a compact representation of deeply nested structures, and Pennylane Operators can easily be converted to and from pytrees. Shortcut story: https://app.shortcut.com/xanaduai/story/63174/datasets-serialization-using-pytrees **Description of the Change:** - Adds a `pytree.serialization` module for converting `PyTreeStructure` objects to and from a JSON representation - Adds a `DatasetPyTree` data attribute type that can store any pytree-compatible type - `DatasetPyTree` is now the default attribute type for `Operator` - The `pytree.unflatten` method disables recording, to prevent deserialized operators from being queued **Benefits:** - Datasets support any Pennylane object with a Pytree representation and serializable metadata - Complex serialization logic in ``DatasetOperator`` can be deprecated - Reduces file size of datasets **Possible Drawbacks:** - Pytree datasets will not be compatible with versions < 0.37 - Type information for metadata is not preserved - `tuple`, `Wires`, and `Shots` objects will all be converted to `list` on a round-trip. `__init__` methods are responsible for ensuring that metadata is converted to the correct type --------- Co-authored-by: albi3ro <chrissie.c.l@gmail.com> Co-authored-by: Christina Lee <christina@xanadu.ai> Co-authored-by: Thomas R. Bromley <49409390+trbromley@users.noreply.github.com> Co-authored-by: Mudit Pandey <mudit.pandey@xanadu.ai> Co-authored-by: Paul Finlay <50180049+doctorperceptron@users.noreply.github.com> Co-authored-by: Diego <67476785+DSGuala@users.noreply.github.com> Co-authored-by: Utkarsh <utkarshazad98@gmail.com> Co-authored-by: Diego <diego_guala@hotmail.com>
**Context:** Catalyst PR #837 (PennyLaneAI/catalyst#837) needs a couple minor updates to the capture module. **Description of the Change:** 1) makes it possible to do `from pennylane.capture import AbstractOperator, AbstractMeasurement, qnode_prim` so we don't have to touch private functions 2) Adds `qnode` as a keyword argument that gets bound to the qnode primitive 3) Makes it so we can capture a sample measurement specified like `qml.sample(wires=1)` **Benefits:** **Possible Drawbacks:** **Related GitHub Issues:** [sc-66703] --------- Co-authored-by: David Wierichs <david.wierichs@xanadu.ai>
…convention changes. (#5853) **Context:** Allow for `qml.qsvt()` to accept phase angles as jax arrays and perform convention conversions. --------- Co-authored-by: David Ittah <dime10@users.noreply.github.com>
**Context:**
Currently the decomposition of a Pauli Rotation with an identity as
Pauli word returns an empty list. I.e,
```python
qml.PauliRot(theta, Pauli_word="I").decomposition() == []
```
**Description of the Change:**
Implement the correct decomposition for this case using
`qml.GlobalPhase`
**Benefits:**
The `TrotterProduct` for Hamiltonians with an identity using
`default.qubit` and `default.lightning` now return the same correct
result.
```python
import pennylane as qml
import numpy as np
symbols = ["H", "H"]
coordinates = np.array([[0.0, 0.0, -0.6614], [0.0, 0.0, 0.6614]])
H, num_qubits = qml.qchem.molecular_hamiltonian(symbols, coordinates, basis="sto-3g")
results = []
for dev_name in ["default.qubit", "lightning.qubit"]:
dev = qml.device(dev_name, wires=num_qubits)
@qml.qnode(dev)
def circuit():
qml.TrotterProduct(H, 1.0)
return qml.state()
results.append(circuit())
qml.math.allclose(*results)
>> True
```
**Related GitHub Issues:**
[#5859]
[sc-66115]
**Context:** Some updates are being made to Catalyst for mid-circuit measurements that are dependent on the changes in this PR. **Description of the Change:** * Change reference in `dynamic_one_shot` to catalyst's `MidCircuitMeasure.out_classical_tracers` to `MidCircuitMeasure.mcm_tracer`. * Change `MidMeasureMP.name` to be more general. This is because `MidCircuitMeasure` will soon inherit `MidMeasureMP`, and this change makes more sense than adding a `name` property to `MidCircuitMeasure`. **Benefits:** **Possible Drawbacks:** **Related GitHub Issues:**
**Context:** This is a follow up to this [PR](#5789) which added detailed documentation for the new PennyLane Debugger feature. --------- Co-authored-by: Astral Cai <astral.cai@xanadu.ai> Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: Yushao Chen (Jerry) <chenys13@outlook.com> Co-authored-by: Christina Lee <chrissie.c.l@gmail.com> Co-authored-by: Mudit Pandey <mudit.pandey@xanadu.ai> Co-authored-by: Thomas R. Bromley <49409390+trbromley@users.noreply.github.com> Co-authored-by: soranjh <40344468+soranjh@users.noreply.github.com>
**Context:** When creating an initial MPS from a dense state vector, the order of wires was not preserved since we were not providing the `wire_order` parameter to `state_vector`. **Description of the Change:** As above. **Benefits:** If the MPS is created from a dense state vector, now the order of wires is preserved. **Possible Drawbacks:** None that I can think of. **Related GitHub Issues:** None. **Related ShortCut Stories**: [sc-66787]
The objective of this PR is to reorganize the changelog into sections. Please don't worry about the content in each bullet point, we can polish that in follow up PRs.
mudit2812
added a commit
that referenced
this pull request
Jul 26, 2024
**Context:** As name says. Gradient workflows no longer raise errors after the merge of #5791 , but their correctness is yet to be verified. **Description of the Change:** * Updated casting rules in `dynamic_one_shot`'s processing function for tensorflow. * For the changes to be fully integrated, the way that the interface is passed around when calling a QNode needed to be changed, so the following changes were made: * `QNode` has updated behaviour for how `mcm_config` is used during execution. In `QNode._execution_component`, a copy of `self.execute_kwargs["mcm_config"]` is the source of truth, and in `qml.execute`, `config.mcm_config` is the source of truth. * Added a private `pad-invalid-samples` `postselect_mode`. The `postselect_mode` is switched to this automatically in `qml.execute` if executing with jax and shots and `postselect_mode == "hw-like"`. This way we standardize how the MCM transforms determine if jax is being used. * Updates to `capture` module to accommodate the above changes. **Benefits:** * `dynamic_one_shot` doesn't cast to interfaces inside the ML boundary * `dynamic_one_shot` works with tensorflow * Expanded tests **Possible Drawbacks:** **Related GitHub Issues:** Fixes #5736, #5710 Duplicate of #5861 which was closed due to release branch merge stuff. --------- Co-authored-by: Jay Soni <jbsoni@uwaterloo.ca> Co-authored-by: Astral Cai <astral.cai@xanadu.ai> Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com> Co-authored-by: Yushao Chen (Jerry) <chenys13@outlook.com> Co-authored-by: Christina Lee <chrissie.c.l@gmail.com> Co-authored-by: Thomas R. Bromley <49409390+trbromley@users.noreply.github.com> Co-authored-by: soranjh <40344468+soranjh@users.noreply.github.com> Co-authored-by: Pietropaolo Frisoni <pietropaolo.frisoni@xanadu.ai> Co-authored-by: Ahmed Darwish <exclass9.24@gmail.com> Co-authored-by: Utkarsh <utkarshazad98@gmail.com> Co-authored-by: David Wierichs <david.wierichs@xanadu.ai> Co-authored-by: Christina Lee <christina@xanadu.ai> Co-authored-by: Mikhail Andrenkov <mikhail@xanadu.ai> Co-authored-by: Diego <67476785+DSGuala@users.noreply.github.com> Co-authored-by: Josh Izaac <josh146@gmail.com> Co-authored-by: Diego <diego_guala@hotmail.com> Co-authored-by: Vincent Michaud-Rioux <vincentm@nanoacademic.com> Co-authored-by: lillian542 <38584660+lillian542@users.noreply.github.com> Co-authored-by: Jack Brown <jack@xanadu.ai> Co-authored-by: Paul Finlay <50180049+doctorperceptron@users.noreply.github.com> Co-authored-by: David Ittah <dime10@users.noreply.github.com> Co-authored-by: Cristian Emiliano Godinez Ramirez <57567043+EmilianoG-byte@users.noreply.github.com> Co-authored-by: Vincent Michaud-Rioux <vincent.michaud-rioux@xanadu.ai>
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.
Context:
As name says. Gradient workflows no longer raise errors after the merge of #5791 , but their correctness is yet to be verified.
Description of the Change:
Benefits:
Possible Drawbacks:
Related GitHub Issues:
Fixes #5736, #5710