Add rng and prng_key to get_final_state, apply_operation
#5337
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
|
[sc-58258] |
albi3ro
reviewed
Mar 7, 2024
vincentmr
added a commit
that referenced
this pull request
Apr 21, 2024
### Before submitting
Please complete the following checklist when submitting a PR:
- [x] All new features must include a unit test.
If you've fixed a bug or added code that should be tested, add a test to
the
test directory!
- [x] All new functions and code must be clearly commented and
documented.
If you do make documentation changes, make sure that the docs build and
render correctly by running `make docs`.
- [x] Ensure that the test suite passes, by running `make test`.
- [x] Add a new entry to the `doc/releases/changelog-dev.md` file,
summarizing the
change, and including a link back to the PR.
- [x] The PennyLane source code conforms to
[PEP8 standards](https://www.python.org/dev/peps/pep-0008/).
We check all of our code against [Pylint](https://www.pylint.org/).
To lint modified files, simply `pip install pylint`, and then
run `pylint pennylane/path/to/file.py`.
When all the above are checked, delete everything above the dashed
line and fill in the pull request template.
------------------------------------------------------------------------------------------------------------
**Context:**
JAX RNG keys should not be reused (in production) according to the JAX
documentation. The current usage only split the keys at the lowest level
(`_sample_state_jax`). This could cause unwanted correlations. For
example, we reuse the device key for each tape in a batch of tapes in
`default.qubit`.
**Description of the Change:**
Add functions and methods to split RNG keys. Split RNG keys to avoid
reusing keys.
**Benefits:**
**Possible Drawbacks:**
**Related GitHub Issues:**
#5337
[sc-61435]
|
Just testing out updating the durations file again. Before updating, this CI run had the following core test durations: |
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## master #5337 +/- ##
==========================================
- Coverage 99.69% 99.69% -0.01%
==========================================
Files 410 412 +2
Lines 38278 38188 -90
==========================================
- Hits 38161 38070 -91
- Misses 117 118 +1 ☔ View full report in Codecov by Sentry. |
|
The new timings are: 11, 22, 21, 21, 24 minutes. |
albi3ro
reviewed
Apr 25, 2024
albi3ro
reviewed
Apr 25, 2024
lillian542
reviewed
Apr 25, 2024
lillian542
approved these changes
Apr 26, 2024
albi3ro
approved these changes
Apr 26, 2024
mudit2812
added
the
merge-ready ✔️
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:
The seed provided to
DefaultQubitduring initialization is only used for measurements. However, we now have uses for it while applying operations, specifically with applying mid-circuit measurements withdynamic_one_shot, and updating the number of shots with postselection withdefer_measurements. This PR updatesdevices.qubit.simulate.get_final_stateanddevices.qubit.apply_operationto use the RNG/PRNGKey provided during device initialization for stochastic behaviour.Description of the Change:
rngandprng_keyarguments toget_final_stateand theMidMeasureMPdispatcher ofapply_operation.appy_operationnow uses the rng/PRNGKey to sample the measurement outcome.rngandprng_keyarguments to_postselection_postprocess. This function is used for renormalizing the state after aProjectoris applied to it, as well as to update the number of shots if provided. This function now uses the rng/PRNGKey to update shots using the binomial distribution.default.qubitnow set a seed on the device. With this change, I also removedflakyfrom all the native MCM tests and instead increased the number of shots for the tests that were still failing.get_final_stateandapply_operationto accept**execution_kwargsarguments, which hiderng,prng_key,interface, andmid_measurements. All of these arguments impact execution outcomes, hence why they were collected. This also means that all of those arguments are now strictly keyword arguments and can no longer be passed to the functions as positional arguments.Benefits:
Possible Drawbacks:
Related GitHub Issues: