Add support for pennylane circuits #836

rmlarose · 2021-07-23T15:17:32Z

Description

Checklist

Check off the following once complete (or if not applicable) after opening the PR. The PR will be reviewed once this checklist is complete and all tests are passing.

I added unit tests for new code.
I used type hints in function signatures.
I used Google-style docstrings for functions.
I updated the documentation where relevant.

If some items remain, you can mark this a draft pull request.

Tips

If the validation check fails:
1. Run make check-types (from the root directory of the repository) and fix any mypy errors.
2. Run make check-style and fix any flake8 errors.
3. Run make format to format your code with the black autoformatter.
For more information, check the Mitiq style guidelines.
Write "Fixes #XYZ" in the description if this PR fixes Issue #XYZ.

rmlarose · 2021-08-01T18:44:04Z

mitiq/interface/mitiq_pennylane/tests/test_conversions_pennylane.py

+def test_to_from_pennylane():
+    circuit = cirq.testing.random_circuit(
+        qubits=4, n_moments=2, op_density=1, random_state=1
+    )
+
+    converted = from_pennylane(to_pennylane(circuit))
+
+    print(circuit)
+    print(converted)
+
+    assert cirq.testing.assert_allclose_up_to_global_phase(
+        cirq.unitary(converted), cirq.unitary(circuit), atol=1e-7,
+    )


Round-trip conversions not preserving qubit order. Output of this test:

┌──┐ 0: ─────×──────── │ 1: ────×┼────H─── ││ 2: ────×┼────@─── │ │ 3: ─────×────@─── └──┘ ┌─────────┐ q_0: ───×────H────────────M('c_0')─── │ q_1: ───×────@────────────M('c_1')─── │ q_2: ───×────┼M('c_2')─────────────── │ │ q_3: ───×────@────────────M('c_3')─── └─────────┘

Hey @trbromley is it possible to reorder the wires of a quantum tape? (Long story motivating this question below.)

Input Cirq circuit:

┌──┐ 0: ─────×──────── │ 1: ────×┼────H─── ││ 2: ────×┼────@─── │ │ 3: ─────×────@─── └──┘

Resulting tape from tape = to_pennylane(circuit):

1: ──╭SWAP───H──┤ 2: ──╰SWAP──╭C──┤ 0: ──╭SWAP──│───┤ 3: ──╰SWAP──╰Z──┤

Note this is the same as above with permuted qubit order. Namely (0, 1, 2, 3) -> (1, 2, 0, 3).

print(tape.to_openqasm(rotations=True))

OPENQASM 2.0; include "qelib1.inc"; qreg q[4]; creg c[4]; swap q[0],q[1]; h q[0]; swap q[2],q[3]; cz q[1],q[3]; measure q[0] -> c[0]; measure q[1] -> c[1]; measure q[2] -> c[2]; measure q[3] -> c[3];

Note the indices of the qubits in the QASM program do not correspond to the indices of the wires in the tape.

(Aside from main point about qubit order, but there are also added measurements that are not in the tape.)

When converting back to Cirq from this QASM, the qubit order is the same as in the QASM program, so is permuted:

┌─────────┐ q_0: ───×────H────────────M('c_0')─── │ q_1: ───×────@────────────M('c_1')─── │ q_2: ───×────┼M('c_2')─────────────── │ │ q_3: ───×────@────────────M('c_3')─── └─────────┘

This can be fixed by reordering the wires in step 2, hence the question.

Hey @rmlarose! Good spot, thanks!

It looks like the tape in step 2 is ok, indeed you can specify the wire order when drawing the tape: tape.draw(wire_order=qml.wires.Wires([0, 1, 2, 3])).

However, it looks like an issue in tape.to_openqasm, since the wires in the QASM script are matched to the order specified by tape.wires = <Wires = [1, 2, 0, 3]>. A quick hack fix is to do:

tape.to_openqasm(wires=sorted(tape.wires))

Overall this looks like a bug/unexpected behaviour and I'll discuss with the team.

Sounds good, thanks! This is currently the main blocker for me. I'm fine with this suggested fix and the requirement that the tape's wires must be sortable. (Which isn't the case with wires labeled [0, "a"], for example.)

Thanks for adding in the fix with pre-sorting the wires! Regarding the follow up comment on the PL-issue, is that a blocker for this PR? 🤔

trbromley

Thanks @rmlarose and sorry for being late to look at this!

The changes look great and thanks for making a start on the tests. My main question is whether we want rotations=True when converting to QASM from a tape. The positive is that we don't need to worry about the observable being measured, but a potential negative is that we might end up folding the rotation gates.

The current approach also loses the PL return type (e.g., measure expval, var, probs). The only way I can think to support that is with rotations=False and adding additional arguments to to_pennylane to persist the output measurement, but I can understand why it's preferable to have just the circuit as an input in to_pennylane.

It's also nontrivial to support custom wire labels (e.g., ["alice", "bob"] rather than integers). 🤔

The simplest solution is probably to restrict the PL circuits that can be mitigated. For example, we could raise an error if the input tape has custom wire labels or non-expval return types. If this makes sense, I'll be happy to make a PR from a fork.

mitiq/interface/mitiq_pennylane/conversions.py

trbromley · 2021-08-04T21:04:19Z

mitiq/interface/mitiq_pennylane/conversions.py

+    Args:
+        tape: Pennylane QuantumTape to convert to a Cirq circuit.
+    """
+    return cirq_from_qasm(tape.to_openqasm(rotations=True))


I'm not sure about the rotations=True part. The rotations diagonalize the measurement so we can just do it in the computational basis.

If we include the rotations at this stage, is it possible that the rotations get folded in he mitigation step? Which may or may not make sense, have to think about it 🤔

It's a good point, thanks for bringing this up! Without regard to folding and just thinking about from_pennylane, it should include the rotations, otherwise the Cirq circuit - which only supports computational basis measurements - will not be equivalent.

However this makes it so that the Pennylane -> Cirq -> Pennylane sequence can start with non-computational basis measurements and end with computational basis measurements, which is a little strange. With hindsight we've realized conversions typically have problems like this, and have handled it in a "first do sloppy conversion, then clean it up" way. (Two examples of cleaning it up are adding back PyQuil declarations and keeping the same register structure in Qiskit.) This is a con of the design in my opinion, but the pro is that as long as you have to_whatever and from_whatever, even if it's (necessarily) a little sloppy, it makes it very easy to add support for whatever.

Anyway, you asked about folding the rotations: Since all other platforms measure in the Z basis we do fold rotations by default. There are options to not fold single-qubit gates, or certain types of gates, so whether or not rotations are there doesn't matter too much from the perspective of folding.

What about having the rotations argument in the function from_pennylane?

It seems that if we consider from_pennylane as a stand-alone function, the natural choice would be rotations=True .

On the other hand, if we also want to restore the full metadata of the tape (return type, observable, etc.) somewhere along the Mitiq worflow, it is probably easier if one can set rotations=False to decouple the bare circuit from the rest.

mitiq/interface/mitiq_pennylane/conversions.py

trbromley · 2021-08-04T21:44:30Z

mitiq/interface/mitiq_pennylane/tests/test_conversions_pennylane.py

+def test_to_from_pennylane():
+    circuit = cirq.testing.random_circuit(
+        qubits=4, n_moments=2, op_density=1, random_state=1
+    )
+
+    converted = from_pennylane(to_pennylane(circuit))
+
+    print(circuit)
+    print(converted)
+
+    assert cirq.testing.assert_allclose_up_to_global_phase(
+        cirq.unitary(converted), cirq.unitary(circuit), atol=1e-7,
+    )


Hey @rmlarose! Good spot, thanks!

It looks like the tape in step 2 is ok, indeed you can specify the wire order when drawing the tape: tape.draw(wire_order=qml.wires.Wires([0, 1, 2, 3])).

However, it looks like an issue in tape.to_openqasm, since the wires in the QASM script are matched to the order specified by tape.wires = <Wires = [1, 2, 0, 3]>. A quick hack fix is to do:

tape.to_openqasm(wires=sorted(tape.wires))

Overall this looks like a bug/unexpected behaviour and I'll discuss with the team.

…tiq_pennylane

codecov · 2021-08-05T18:59:02Z

Codecov Report

Merging #836 (eb143d1) into master (6445c0b) will increase coverage by 0.09%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #836      +/-   ##
==========================================
+ Coverage   97.87%   97.96%   +0.09%     
==========================================
  Files          53       55       +2     
  Lines        2541     2605      +64     
==========================================
+ Hits         2487     2552      +65     
+ Misses         54       53       -1

Impacted Files	Coverage Δ
mitiq/_typing.py	`100.00% <100.00%> (ø)`
mitiq/interface/conversions.py	`99.09% <100.00%> (+0.06%)`	⬆️
mitiq/interface/mitiq_pennylane/__init__.py	`100.00% <100.00%> (ø)`
mitiq/interface/mitiq_pennylane/conversions.py	`100.00% <100.00%> (ø)`
mitiq/__init__.py	`100.00% <0.00%> (ø)`
mitiq/executor/executor.py	`100.00% <0.00%> (ø)`
mitiq/interface/mitiq_cirq/cirq_utils.py	`100.00% <0.00%> (ø)`
mitiq/zne/inference.py	`97.42% <0.00%> (+0.07%)`	⬆️
mitiq/zne/zne.py	`96.66% <0.00%> (+0.11%)`	⬆️
mitiq/observable/observable.py	`100.00% <0.00%> (+1.35%)`	⬆️

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 6445c0b...eb143d1. Read the comment docs.

trbromley

Thanks @rmlarose! This is looking great!

To confirm, the remaining item left to do is to add a check for the return type of the quantum tape in from_pennylane() to ensure that only returning expectation values (for now) is supported? I can help with this, just want to make sure you're on board.

I'd also like to do a few manual trial runs of mitigating PL tapes using this new functionality, to test for any edge cases I can think of. Is there anything remaining to be done in the frontend to get the PL-tape compatible with, for example, execute_with_zne or the higher level folding functionality?

dev_requirements.txt

trbromley · 2021-08-16T18:37:34Z

mitiq/interface/mitiq_pennylane/tests/test_conversions_pennylane.py

+def test_to_from_pennylane():
+    circuit = cirq.testing.random_circuit(
+        qubits=4, n_moments=2, op_density=1, random_state=1
+    )
+
+    converted = from_pennylane(to_pennylane(circuit))
+
+    print(circuit)
+    print(converted)
+
+    assert cirq.testing.assert_allclose_up_to_global_phase(
+        cirq.unitary(converted), cirq.unitary(circuit), atol=1e-7,
+    )


Thanks for adding in the fix with pre-sorting the wires! Regarding the follow up comment on the PL-issue, is that a blocker for this PR? 🤔

rmlarose · 2021-08-16T20:09:30Z

Thanks @trbromley. I just pushed code adding Pennylane tapes to mitiq.QPROGRAM and the top-level to/from conversions. It's still a bit rough around the edges, and needs more tests.

Thanks for adding in the fix with pre-sorting the wires! Regarding the follow up comment on the PL-issue, is that a blocker for this PR? 🤔

There are some failing tests with PEC, so I would consider that blocking, yes.

To confirm, the remaining item left to do is to add a check for the return type of the quantum tape in from_pennylane() to ensure that only returning expectation values (for now) is supported? I can help with this, just want to make sure you're on board.

I'm actually not sure what you mean by this, sorry. Can you elaborate?

Is there anything remaining to be done in the frontend to get the PL-tape compatible with, for example, execute_with_zne or the higher level folding functionality?

Probably - An end-to-end example will tell. With the latest code we can kinda sorta fold gates in Pennylane tapes:

import numpy as np

import pennylane as qml
from mitiq import zne

with qml.tape.QuantumTape() as tape:
    qml.RX(np.pi / 2, wires=[0])
    qml.CNOT(wires=[0, 1])
    qml.expval(qml.PauliZ(wires=[0]))
    qml.expval(qml.PauliZ(wires=[1]))

print(tape.draw())

scaled_tape = zne.scaling.fold_all(tape, scale_factor=3)
print(scaled_tape.draw())

 0: ──RX(1.57)──╭C──┤ ⟨Z⟩ 
 1: ────────────╰X──┤ ⟨Z⟩ 


 0: ──RX(1.57)──RX(-1.57)──RX(1.57)──╭C──╭C──╭C──┤  
 1: ─────────────────────────────────╰X──╰X──╰X──┤

Note the ⟨Z⟩s are removed in the scaled_tape. This would probably cause issues in an end-to-end example.

trbromley

Thanks @rmlarose!

To confirm, the remaining item left to do is to add a check for the return type of the quantum tape in from_pennylane() to ensure that only returning expectation values (for now) is supported? I can help with this, just want to make sure you're on board.

I'm actually not sure what you mean by this, sorry. Can you elaborate?

The suggestions I've left in this review should fix things. In PennyLane we can have other return types, for example:

with qml.tape.QuantumTape() as tape:
    qml.RX(0.2, wires=0)
    qml.RY(0.6, wires=1)
    qml.CNOT(wires=[0, 1])
    qml.var(qml.PauliY(0))

This would give the variance of PauliY. I think for this PR it's sufficient to support just expectation values.

There are some failing tests with PEC, so I would consider that blocking, yes.

This should be fixed in PennyLaneAI/pennylane#1559 - though this would not make it to a release version of PennyLane until September 🤔

mitiq/interface/mitiq_pennylane/conversions.py

mitiq/interface/mitiq_pennylane/tests/test_conversions_pennylane.py

nathanshammah · 2021-09-01T14:08:24Z

The validate CI fails seems fixable with black. The test/test fail seems do to an AssertionError .

crazy4pi314 · 2021-09-01T20:55:03Z

The validate CI fails seems fixable with black. The test/test fail seems do to an AssertionError .

@nathanshammah should have fixed this and merged in master to continue testing

Co-authored-by: Tom Bromley <49409390+trbromley@users.noreply.github.com>

…ne.py Co-authored-by: Tom Bromley <49409390+trbromley@users.noreply.github.com>

Co-authored-by: Tom Bromley <49409390+trbromley@users.noreply.github.com>

andreamari · 2021-09-20T16:57:03Z

Hello everyone, a quick update.

There was a remaining failing test due to a problem related to conversions of identity gates.
So I updated the failing test from an exact circuit comparison to a matrix comparison:

mitiq/mitiq/interface/mitiq_pennylane/tests/test_conversions_pennylane.py

Line 112 in e82132d

# TODO: test circuit equality after Identity operation will be added

The reason is that the Identity operation doesn't exist in PennyLane at the moment although there is a plan to add it (PennyLaneAI/pennylane#1632).
See, also PennyLaneAI/pennylane-qiskit#153.

Once those issues will be fixed, we can improve the test asserting gate equality instead of matrix equality.

There may be other subtle edge cases that I didn't notice. So feel free to add more tests.

A thing to decide before merging is what kind of tape we want to explicitly support. Do we want to raise an error each time a tape contains other things that are not gates (measurements, expvals, etc..)? This is a sub-optimal but probably safer solution.
If I am not wrong, @trbromley suggested something similar during the last PL community call.

nathanshammah · 2021-09-23T13:30:57Z

About @andreamari's proposal:

A thing to decide before merging is what kind of tape we want to explicitly support. Do we want to raise an error each time a tape contains other things that are not gates (measurements, expvals, etc..)? This is a sub-optimal but probably safer solution.

I am in favor of this. We can extend functionality later. We could then add an example with basic integration. @trbromley what do you think?

nathanshammah · 2021-09-28T13:32:26Z

I resolved the conflict updating the requirements to include the latest braket update on master branch.

trbromley

Thanks everyone! This looks great.

I agree with @andreamari's suggestion to restrict the input tape to not have any measurements at this stage.

I have added a few suggestions, including an integration test that should give a deeper confirmation that the conversion is working as expected.

dev_requirements.txt

trbromley · 2021-09-29T20:37:13Z

mitiq/interface/mitiq_pennylane/conversions.py

+    Returns:
+        Mitiq circuit representation equivalent to the input QuantumTape.
+    """
+    try:


@WrathfulSpatula, agreed! We should probably check that the input wires are contiguous. As @rmlarose points out, the try/except is for the case of custom wire labels and I believe is still valid.

trbromley · 2021-09-29T20:40:30Z

mitiq/interface/mitiq_pennylane/conversions.py

+    # tape.to_openqasm always introduces measurements that we remove
+    measurements = _pop_measurements(output)
+    # Remove final empty moment if present
+    if measurements and output[-1] == Moment():
+        return output[:-1]


Thanks @andreamari! An alternative approach might be to use tape.to_openqasm(rotations=True, wires=wires, measure_all=False), which we added in PennyLaneAI/pennylane#1559. This would need to be complemented by ensuring that tape.measurements is empty (as suggested).

The above would also require PL >= 0.18. So maybe your solution is perfect as is!

+1 for tape.to_openqasm(rotations=True, wires=wires, measure_all=False)

Applied in this commit: eb143d1 . It works!

I also explicitly added pennylane>=0.18 in dev_requirements.txt to ensure previous versions of PL are updated in local environments.

mitiq/interface/mitiq_pennylane/conversions.py

mitiq/interface/mitiq_pennylane/tests/test_conversions_pennylane.py

mitiq/interface/mitiq_pennylane/conversions.py

mitiq/interface/mitiq_pennylane/tests/test_conversions_pennylane.py

Co-authored-by: Tom Bromley <49409390+trbromley@users.noreply.github.com>

andreamari · 2021-10-01T12:47:09Z

Thanks for all the great suggestions, @trbromley !

Side note: I am impressed that, after directly applying all your complex GitHub suggestions, tests passed out of the box! 😃

I am approving this PR, but I would wait for a double check by other Mitiq people before merging.

trbromley

Thanks @andreamari for the latest updates and to everyone for contributing to this! It looks good to go and I'm excited to have this integration 🚀

Regarding updating the documentation (e.g. adding to here), is that best done in a follow up PR? I'd think so, it's nice to get this PR across the line first.

trbromley · 2021-10-01T20:18:57Z

dev_requirements.txt

@@ -3,6 +3,8 @@ qiskit-terra~=0.18.2
 qiskit-aer~=0.8.2
 qiskit-ibmq-provider~=0.16.0
 pyquil~=2.28.0
+pennylane-qiskit~=0.18.0
+pennylane>=0.18


Do you specify anywhere in the docs a minimum version for soft dependencies?

@crazy4pi314, @rmlarose, @nathanshammah what do you think?

In principle it would be nice to add info about minimum versions but there are some downsides:

It can easily become outdated.

In general, we actually don't know what are the minimum versions since we only test the specific versions pinned in dev_requirements.txt.

Andrea, I agree on the fact that it would be nice to show soft dependencies in the documentation.

It is also true as you say that with the dependabot in place, updating dependencies quite often, it would add an overhead to test the minimum versions working. I don't know what is a best practice here.

In my opinion we should keep the current dev requirements consistent with pinning to the minor (pennylane~=0.18); we could add to the docs a citation of the requirements and dev requirements, , e.g., scraping those files.

crazy4pi314 · 2021-10-07T19:47:31Z

I'll merge this now, and make sure we have an issue for user docs on this! Grats everyone 🎉

andreamari · 2021-10-08T10:26:19Z

Even if tests were passing in this PR, after merging on master tests are failing for some pip installation problem.
~~I would propose clicking on the revert button above to undo the changes until we figure out how to fix the problem.~~
Problem solved in #976.

Take 1.

c3c68f3

rmlarose marked this pull request as draft July 23, 2021 15:17

rmlarose added 3 commits July 30, 2021 20:41

CP.

a9850f9

Remove pennylane-qiskit from dev requirements.

cb55d2f

Init conversions and (failing) tests.

3129dd0

rmlarose commented Aug 1, 2021

View reviewed changes

Format.

9bf9e7b

trbromley reviewed Aug 4, 2021

View reviewed changes

rmlarose added 2 commits August 5, 2021 00:54

Merge branch 'master' of https://github.com/unitaryfund/mitiq into mi…

6d88d08

…tiq_pennylane

Add pennylane to dev requirements.

b80a907

trbromley mentioned this pull request Aug 5, 2021

[BUG] Unexpected behaviour when using tape.to_openqasm() PennyLaneAI/pennylane#1500

Open

1 task

rmlarose added 4 commits August 5, 2021 14:23

Pennylane-qiskit.

6094e2e

Sort wires.

04d709b

Fix test.

f28570d

More tests.

c8435c6

UnsupportedQuantumTapeError and tests.

41b0d9c

trbromley reviewed Aug 16, 2021

View reviewed changes

Add pennylane to QPROGRAM and to/from conversions. Needs tests.

4c54fff

Merge branch 'master' into mitiq_pennylane

d7d5ab2

trbromley reviewed Aug 18, 2021

View reviewed changes

mitiq/interface/mitiq_pennylane/conversions.py Show resolved Hide resolved

mitiq/interface/mitiq_pennylane/conversions.py Show resolved Hide resolved

mitiq/interface/mitiq_pennylane/tests/test_conversions_pennylane.py Show resolved Hide resolved

crazy4pi314 self-assigned this Aug 20, 2021

Merge branch 'master' into mitiq_pennylane, fix black

528f8e0

crazy4pi314 and others added 4 commits September 8, 2021 10:45

Update mitiq/interface/mitiq_pennylane/conversions.py

1387a15

Co-authored-by: Tom Bromley <49409390+trbromley@users.noreply.github.com>

Update mitiq/interface/mitiq_pennylane/tests/test_conversions_pennyla…

b897c90

…ne.py Co-authored-by: Tom Bromley <49409390+trbromley@users.noreply.github.com>

Update mitiq/interface/mitiq_pennylane/conversions.py

af8144c

Co-authored-by: Tom Bromley <49409390+trbromley@users.noreply.github.com>

Apply suggestions from code review

af85499

Co-authored-by: Tom Bromley <49409390+trbromley@users.noreply.github.com>

andreamari added 2 commits September 23, 2021 11:13

check wires are contiguous

7283e56

fix format

a082169

Merge branch 'master' into mitiq_pennylane

ba65d53

trbromley reviewed Sep 29, 2021

View reviewed changes

andreamari and others added 11 commits October 1, 2021 12:08

Update dev_requirements.txt

2be7f0b

Co-authored-by: Tom Bromley <49409390+trbromley@users.noreply.github.com>

ensure there are no measurements

2ac0484

Co-authored-by: Tom Bromley <49409390+trbromley@users.noreply.github.com>

remove old expval tests

6869ef6

Co-authored-by: Tom Bromley <49409390+trbromley@users.noreply.github.com>

define unsupported operations

e11e62f

Co-authored-by: Tom Bromley <49409390+trbromley@users.noreply.github.com>

expand tape to supported ops

aa7ecf6

Co-authored-by: Tom Bromley <49409390+trbromley@users.noreply.github.com>

add integration test

b81fec7

Co-authored-by: Tom Bromley <49409390+trbromley@users.noreply.github.com>

import Expectation

747f8c4

Co-authored-by: Tom Bromley <49409390+trbromley@users.noreply.github.com>

update error message in test

18b722e

Co-authored-by: Tom Bromley <49409390+trbromley@users.noreply.github.com>

black

badf53f

fix flake

57a2a6e

make use of measure_all=False option

eb143d1

andreamari approved these changes Oct 1, 2021

View reviewed changes

trbromley approved these changes Oct 1, 2021

View reviewed changes

andreamari mentioned this pull request Oct 7, 2021

Add PennyLane to frontend table in the readme #973

Merged

crazy4pi314 merged commit deec3ef into master Oct 7, 2021

crazy4pi314 mentioned this pull request Oct 8, 2021

User documentation for the PennyLane integration #974

Closed

rmlarose deleted the mitiq_pennylane branch October 11, 2021 16:17

rmlarose mentioned this pull request Oct 11, 2021

Pennylane integration #817

Closed

sassan72 mentioned this pull request Oct 19, 2021

mitiq can't support pennylane. #990

Closed

trbromley mentioned this pull request Oct 21, 2021

[WIP] Add mitigation support using mitiq backend PennyLaneAI/pennylane#1783

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add support for pennylane circuits #836

Add support for pennylane circuits #836

rmlarose commented Jul 23, 2021

rmlarose Aug 1, 2021

rmlarose Aug 2, 2021

trbromley Aug 4, 2021

rmlarose Aug 5, 2021

trbromley Aug 16, 2021

trbromley left a comment •

edited

Loading

trbromley Aug 4, 2021

rmlarose Aug 5, 2021

andreamari Aug 6, 2021

trbromley Aug 4, 2021

codecov bot commented Aug 5, 2021 •

edited

Loading

trbromley left a comment

trbromley Aug 16, 2021

rmlarose commented Aug 16, 2021

trbromley left a comment

nathanshammah commented Sep 1, 2021

crazy4pi314 commented Sep 1, 2021 •

edited

Loading

andreamari commented Sep 20, 2021 •

edited

Loading

nathanshammah commented Sep 23, 2021 •

edited

Loading

nathanshammah commented Sep 28, 2021

trbromley left a comment

trbromley Sep 29, 2021

trbromley Sep 29, 2021 •

edited

Loading

andreamari Oct 1, 2021

andreamari Oct 1, 2021 •

edited

Loading

andreamari commented Oct 1, 2021

trbromley left a comment

trbromley Oct 1, 2021

andreamari Oct 5, 2021

nathanshammah Oct 7, 2021

crazy4pi314 commented Oct 7, 2021

andreamari commented Oct 8, 2021 •

edited

Loading

Add support for pennylane circuits #836

Add support for pennylane circuits #836

Conversation

rmlarose commented Jul 23, 2021

Description

Checklist

Tips

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

trbromley left a comment • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

codecov bot commented Aug 5, 2021 • edited Loading

Codecov Report

trbromley left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rmlarose commented Aug 16, 2021

trbromley left a comment

Choose a reason for hiding this comment

nathanshammah commented Sep 1, 2021

crazy4pi314 commented Sep 1, 2021 • edited Loading

andreamari commented Sep 20, 2021 • edited Loading

nathanshammah commented Sep 23, 2021 • edited Loading

nathanshammah commented Sep 28, 2021

trbromley left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

trbromley Sep 29, 2021 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

andreamari Oct 1, 2021 • edited Loading

Choose a reason for hiding this comment

andreamari commented Oct 1, 2021

trbromley left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

crazy4pi314 commented Oct 7, 2021

andreamari commented Oct 8, 2021 • edited Loading

trbromley left a comment •

edited

Loading

codecov bot commented Aug 5, 2021 •

edited

Loading

crazy4pi314 commented Sep 1, 2021 •

edited

Loading

andreamari commented Sep 20, 2021 •

edited

Loading

nathanshammah commented Sep 23, 2021 •

edited

Loading

trbromley Sep 29, 2021 •

edited

Loading

andreamari Oct 1, 2021 •

edited

Loading

andreamari commented Oct 8, 2021 •

edited

Loading