Clifford simultaneous filtered rb #412

vodovozovaliza · 2023-06-20T13:27:42Z

(Juan): This will need qiboteam/qibo#1015 to save and load circuits properly

Checklist:

Reviewers confirm new code works as expected.
Tests are passing.
Coverage does not decrease.
Documentation is updated.

codecov · 2023-06-20T13:30:10Z

Codecov Report

Merging #412 (76f40c0) into main (e8774fd) will increase coverage by 0.12%.
Report is 50 commits behind head on main.
The diff coverage is 96.91%.

@@            Coverage Diff             @@
##             main     #412      +/-   ##
==========================================
+ Coverage   95.90%   96.02%   +0.12%     
==========================================
  Files          83       85       +2     
  Lines        5740     5994     +254     
==========================================
+ Hits         5505     5756     +251     
- Misses        235      238       +3

Flag	Coverage Δ
unittests	`96.02% <96.91%> (+0.12%)`	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files	Coverage Δ
src/qibocal/protocols/characterization/__init__.py	`100.00% <100.00%> (ø)`
...s/characterization/randomized_benchmarking/data.py	`100.00% <100.00%> (ø)`
...haracterization/randomized_benchmarking/fitting.py	`100.00% <100.00%> (+6.34%)`	⬆️
...cterization/randomized_benchmarking/noisemodels.py	`100.00% <100.00%> (ø)`
...cterization/randomized_benchmarking/standard_rb.py	`97.35% <100.00%> (-1.33%)`	⬇️
...on/randomized_benchmarking/clifford_filtered_rb.py	`99.29% <99.29%> (ø)`
.../characterization/randomized_benchmarking/utils.py	`98.64% <96.96%> (-1.36%)`	⬇️
...s/characterization/randomized_benchmarking/plot.py	`95.34% <95.34%> (ø)`
...erization/randomized_benchmarking/circuit_tools.py	`82.92% <66.66%> (+0.42%)`	⬆️

... and 1 file with indirect coverage changes

andrea-pasquale · 2023-08-21T10:54:24Z

@vodovozovaliza @wilkensJ what is the status of this PR?

andrea-pasquale · 2023-09-11T07:37:26Z

@ingoroth is someone going to work on this?

ingoroth · 2023-09-13T05:39:38Z

At the moment nobody is actively working on this anymore. @vodovozovaliza might find some time to push this again at some point in the future. Maybe @Jacfomg has capacities to take over.

Jacfomg · 2023-09-13T08:08:55Z

At the moment nobody is actively working on this anymore. @vodovozovaliza might find some time to push this again at some point in the future. Maybe @Jacfomg has capacities to take over.

We can talk about it next week if it's ok

for more information, see https://pre-commit.ci

…bocal into clifford_filtered_rb

for more information, see https://pre-commit.ci

andrea-pasquale · 2023-10-05T10:09:22Z

@Jacfomg feel free to merge main and update the qibo dep in pyproject.toml.

I believe that there is an issue with latest release of qibolab. I'm having troubles also here #539 :(

Jacfomg · 2023-10-05T10:58:06Z

I think those tests should pass once we merge #499 that has the coupler changes needed.

andrea-pasquale

Thanks for the updates @Jacfomg.
A few more comments.
I tried to run the protocol using simulation

Perhaps it is better to move the legend under the plot with a horizontal layout.
Moreover, I'm seeing that in the matrix a color is shown as nan is it normal or am I doing something wrong?

src/qibocal/protocols/characterization/randomized_benchmarking/clifford_filtered_rb.py

andrea-pasquale · 2023-10-20T13:25:51Z

@Jacfomg I believe that if we manage to run at least once on hardware without errors we should be ready to merge this PR.

andrea-pasquale · 2023-10-25T12:50:28Z

src/qibocal/protocols/characterization/randomized_benchmarking/clifford_filtered_rb.py

Just a quick comment here.
Even if the parameters/data/results are the same as the StandardRB could you please add a separate class for each just for readability?

andrea-pasquale · 2023-10-25T12:51:43Z

Executed on hardware: http://login.qrccluster.com:9000/JdmSqP64RrG-t4WRSd9zhw==/

andrea-pasquale · 2023-10-25T12:51:57Z

Could you also fix conflicts?

Edoardo-Pedicillo

Documentation is missing for some functions; it would be great to add it.

Edoardo-Pedicillo · 2023-11-06T12:22:23Z

src/qibocal/protocols/characterization/randomized_benchmarking/circuit_tools.py

+    if isinstance(new_layer, Gate):
+        nqubits = max(new_layer.qubits) + 1
+    elif all(isinstance(gate, Gate) for gate in new_layer):
+        nqubits = max(max(gate.qubits) for gate in new_layer) + 1


If new_layer is always an iterable, you can reduce it to one if.

Edoardo-Pedicillo · 2023-11-06T12:26:46Z

src/qibocal/protocols/characterization/randomized_benchmarking/circuit_tools.py

+    elif isinstance(new_layer, Circuit):
+        nqubits = new_layer.nqubits
+    else:


Why does layer_gen return list of Gates or Circuit? A list of Gates can be seen as a circuit and vice versa.

I agree. I think the original functionality of passing either a single Gate or a Circuit is sufficient. I guess the motivation was to be able to pass a layer of parallel single qubit gates. But this can also be cast to Circuit when passed.

Edoardo-Pedicillo · 2023-11-06T12:32:08Z

src/qibocal/protocols/characterization/randomized_benchmarking/circuit_tools.py

+    else:
+        raise_error(
+            TypeError,
+            f"layer_gen must return type Circuit or Gate, but it is type {type(new_layer)}.",
+        )
+


Suggested change

else:

raise_error(

TypeError,

f"layer_gen must return type Circuit or Gate, but it is type {type(new_layer)}.",

)

Is there a particular reason for raising this error? We're all adults here; I'm guessing the code will crash if the wrong type is used.

Edoardo-Pedicillo · 2023-11-06T12:35:14Z

src/qibocal/protocols/characterization/randomized_benchmarking/circuit_tools.py

+    if not isinstance(depth, int) or depth < 0:
+        raise_error(ValueError, f"Depth: {depth}, must be type int and >= 0.")
+
+    # Generate a layer to get nqubits.


I don't understand the first part of this function, if you need to know the number of qubits, you can add it as an input.

In principle there is no need to pass nqubits as it is implicit in the output of layer_gen.__call__(). But I don't like that now the first call of the generator is not added to the circuit. This makes it harder when one actually has to reproduce the result of a generator, say by restoring the rng state.

Edoardo-Pedicillo · 2023-11-06T13:03:22Z

src/qibocal/protocols/characterization/randomized_benchmarking/clifford_filtered_rb.py

+        if platform and platform.name != "dummy":
+            raise_error(
+                NotImplementedError,
+                f"Backend qibolab ({platform}) does not perform noise models simulation.",
+            )


Can we skip this if and set directly the numpy backend?

Edoardo-Pedicillo · 2023-11-06T16:51:04Z

src/qibocal/protocols/characterization/randomized_benchmarking/clifford_filtered_rb.py

+        # Compute fitting uncertainties
+        if len(popt_estimates):
+            perr = data_uncertainties(popt_estimates, uncertainties, data_median=popt)
+            perr = perr.T if perr is not None else (0,) * len(popt)


Suggested change

perr = perr.T if perr is not None else (0,) * len(popt)

perr = perr.T if perr else (0,) * len(popt)

Edoardo-Pedicillo · 2023-11-06T16:54:52Z

src/qibocal/protocols/characterization/randomized_benchmarking/clifford_filtered_rb.py

+    # TODO: If fit is None: loop to separate plotting for fitting in qq
+
+    # TODO: fit doesn't get loaded on acquisition tests


There are a lot of TODOS, I suggest to collect them in an issue.

Edoardo-Pedicillo · 2023-11-06T16:55:29Z

src/qibocal/protocols/characterization/randomized_benchmarking/clifford_filtered_rb.py

+    # TODO: This mess as well ...
+    # meta_data = deepcopy(data.attrs)
+    # meta_data.pop("depths", None)
+    # if not meta_data["noise_model"]:
+    #     meta_data.pop("noise_model")
+    #     meta_data.pop("noise_params")
+    # elif meta_data.get("noise_params", None) is not None:
+    #     meta_data["noise_params"] = np.round(meta_data["noise_params"], 3)
+
+    # table_str = "".join(
+    #     [f" | {key}: {value}<br>" for key, value in {**meta_data, **rb_params}.items()]
+    # )
+
+    # table_str = table_html(table_dict("Something"))


Dead code, delete it

Edoardo-Pedicillo · 2023-11-06T16:58:56Z

src/qibocal/protocols/characterization/randomized_benchmarking/data.py

+        if gate["name"] == "rz":
+            gatelist.append(
+                gates.RZ(gate["_target_qubits"][0], gate["init_kwargs"]["theta"])
+            )
+        if gate["name"] == "rx":
+            gatelist.append(
+                gates.RX(gate["_target_qubits"][0], gate["init_kwargs"]["theta"])
+            )
+        if gate["name"] == "ry":
+            gatelist.append(
+                gates.RY(gate["_target_qubits"][0], gate["init_kwargs"]["theta"])
+            )


This lines can be collected in a for loop using getattr

Edoardo-Pedicillo · 2023-11-06T16:59:16Z

src/qibocal/protocols/characterization/randomized_benchmarking/data.py

+        if gate["name"] == "z":
+            gatelist.append(gates.Z(gate["_target_qubits"][0]))
+        if gate["name"] == "x":
+            gatelist.append(gates.X(gate["_target_qubits"][0]))
+        if gate["name"] == "y":
+            gatelist.append(gates.Y(gate["_target_qubits"][0]))


also this one

Edoardo-Pedicillo · 2024-03-29T09:12:34Z

This PR is outdated, and it seems it will not be concluded, btw the protocol is interesting so I have opened #779

vodovozovaliza added 7 commits June 13, 2023 17:39

Initiall commit

dc917ce

Clifford simultaneous filtered RB

e5d80c3

Merge branch 'liza/rb_bootstrap'

affa5b5

Merge branch 'liza/rb_bootstrap' into clifford_filtered_rb

dc17697

Minor change

d51e258

Merge branch 'liza/rb_bootstrap' into clifford_filtered_rb

174b2a5

Change plot

7d9d83b

vodovozovaliza self-assigned this Jun 20, 2023

vodovozovaliza added 3 commits June 20, 2023 18:51

Add tests

57e81e5

change display | add seed to noise models

d3a6306

Fix coverage

21997f3

Base automatically changed from liza/rb_bootstrap to main June 22, 2023 13:51

vodovozovaliza added 2 commits July 6, 2023 17:56

Fix plots

2186a78

Merge branch 'main' into clifford_filtered_rb

6be3ea5

andrea-pasquale added this to the Qibocal 0.0.5 milestone Aug 31, 2023

Jacfomg and others added 2 commits September 18, 2023 12:32

Merge branch 'main' into clifford_filtered_rb

95aba30

[pre-commit.ci] auto fixes from pre-commit.com hooks

3824b52

for more information, see https://pre-commit.ci

Jacfomg mentioned this pull request Sep 21, 2023

Save gates to JSON qiboteam/qibo#1015

Merged

4 tasks

Jacfomg and others added 7 commits September 21, 2023 11:42

save and load circuits

e2c1194

Merge branch 'clifford_filtered_rb' of https://github.com/qiboteam/qi…

73f7596

…bocal into clifford_filtered_rb

working

93ca115

comment

39353f6

[pre-commit.ci] auto fixes from pre-commit.com hooks

31b1208

for more information, see https://pre-commit.ci

rename json

dff5910

fixes

724e7f2

Jacfomg and others added 2 commits October 5, 2023 12:35

Merge branch 'main' into clifford_filtered_rb

27f3f8a

[pre-commit.ci] auto fixes from pre-commit.com hooks

e521611

for more information, see https://pre-commit.ci

update qibolab

c343bd9

scarrazza modified the milestones: Qibocal 0.0.4, Qibocal 0.0.5 Oct 10, 2023

Merge branch 'main' into clifford_filtered_rb

0d01232

andrea-pasquale removed do not merge Soft warning to avoid merging a PR for reason provided in the PR qibo release required PR requires qibo feature not available in current poetry.lock labels Oct 11, 2023

andrea-pasquale reviewed Oct 12, 2023

View reviewed changes

src/qibocal/protocols/characterization/randomized_benchmarking/clifford_filtered_rb.py Outdated Show resolved Hide resolved

src/qibocal/protocols/characterization/randomized_benchmarking/clifford_filtered_rb.py Show resolved Hide resolved

table

76f40c0

andrea-pasquale added the deploy on hardware label Oct 18, 2023

andrea-pasquale self-requested a review October 18, 2023 09:45

andrea-pasquale reviewed Oct 25, 2023

View reviewed changes

scarrazza modified the milestones: Qibocal 0.0.5, Qibocal 0.0.6 Nov 2, 2023

andrea-pasquale added the rb Randomized benchmarking label Nov 6, 2023

Edoardo-Pedicillo requested changes Nov 6, 2023

View reviewed changes

scarrazza modified the milestones: Qibocal 0.0.6, Qibocal 0.0.7 Nov 29, 2023

andrea-pasquale modified the milestones: Qibocal 0.0.7, Qibocal 0.0.8 Dec 18, 2023

Edoardo-Pedicillo mentioned this pull request Mar 29, 2024

Implement filtered RB #779

Open

Edoardo-Pedicillo closed this Mar 29, 2024

andrea-pasquale deleted the clifford_filtered_rb branch June 10, 2024 08:32

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Clifford simultaneous filtered rb #412

Clifford simultaneous filtered rb #412

vodovozovaliza commented Jun 20, 2023 •

edited by Jacfomg

Loading

codecov bot commented Jun 20, 2023 •

edited

Loading

andrea-pasquale commented Aug 21, 2023

andrea-pasquale commented Sep 11, 2023

ingoroth commented Sep 13, 2023

Jacfomg commented Sep 13, 2023

andrea-pasquale commented Oct 5, 2023

Jacfomg commented Oct 5, 2023

andrea-pasquale left a comment

andrea-pasquale commented Oct 20, 2023

andrea-pasquale Oct 25, 2023

andrea-pasquale commented Oct 25, 2023

andrea-pasquale commented Oct 25, 2023

Edoardo-Pedicillo left a comment

Edoardo-Pedicillo Nov 6, 2023 •

edited

Loading

Edoardo-Pedicillo Nov 6, 2023

ingoroth Nov 7, 2023

Edoardo-Pedicillo Nov 6, 2023

Edoardo-Pedicillo Nov 6, 2023

ingoroth Nov 7, 2023

Edoardo-Pedicillo Nov 6, 2023

Edoardo-Pedicillo Nov 6, 2023

Edoardo-Pedicillo Nov 6, 2023

Edoardo-Pedicillo Nov 6, 2023

Edoardo-Pedicillo Nov 6, 2023

Edoardo-Pedicillo Nov 6, 2023

Edoardo-Pedicillo commented Mar 29, 2024 •

edited

Loading

	perr = perr.T if perr is not None else (0,) * len(popt)
	perr = perr.T if perr else (0,) * len(popt)

		# TODO: If fit is None: loop to separate plotting for fitting in qq

		# TODO: fit doesn't get loaded on acquisition tests

Clifford simultaneous filtered rb #412

Clifford simultaneous filtered rb #412

Conversation

vodovozovaliza commented Jun 20, 2023 • edited by Jacfomg Loading

codecov bot commented Jun 20, 2023 • edited Loading

Codecov Report

andrea-pasquale commented Aug 21, 2023

andrea-pasquale commented Sep 11, 2023

ingoroth commented Sep 13, 2023

Jacfomg commented Sep 13, 2023

andrea-pasquale commented Oct 5, 2023

Jacfomg commented Oct 5, 2023

andrea-pasquale left a comment

Choose a reason for hiding this comment

andrea-pasquale commented Oct 20, 2023

Choose a reason for hiding this comment

andrea-pasquale commented Oct 25, 2023

andrea-pasquale commented Oct 25, 2023

Edoardo-Pedicillo left a comment

Choose a reason for hiding this comment

Edoardo-Pedicillo Nov 6, 2023 • 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

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

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Edoardo-Pedicillo commented Mar 29, 2024 • edited Loading

vodovozovaliza commented Jun 20, 2023 •

edited by Jacfomg

Loading

codecov bot commented Jun 20, 2023 •

edited

Loading

Edoardo-Pedicillo Nov 6, 2023 •

edited

Loading

Edoardo-Pedicillo commented Mar 29, 2024 •

edited

Loading