Skip to main content

An implementation of ILP decoders for quantum error correction.

Project description

ILP Decoder

An implementation of the ILP decoder for circuit-level noise.

Included is a ilp_decoder.ILPDecoder class that can be configured using a stim.DetectorErrorModel and decodes shot data, directly outputting predicted observables (without sinter), as well as a ilp_decoder.ILPSinterDecoder class, which subclasses sinter.Decoder, for interfacing with sinter.

Installation

To install from pypi, run:

pip install ilp_decoder

To install from source, run:

pip install -e .

from the root directory.

Usage

Here is an example of how the decoder can be used directly with Stim:

import stim
import numpy as np
from ilp_decoder import ILPDecoder

num_shots = 100
d = 3
p = 0.001
circuit = stim.Circuit.generated(
    "surface_code:rotated_memory_x",
    rounds=d,
    distance=d,
    before_round_data_depolarization=p,
    before_measure_flip_probability=p,
    after_reset_flip_probability=p,
    after_clifford_depolarization=p
)

sampler = circuit.compile_detector_sampler()
shots, observables = sampler.sample(num_shots, separate_observables=True)

decoder = ILPDecoder.from_circuit(circuit)

predicted_observables = decoder.decode_batch(shots)
num_mistakes = np.sum(np.any(predicted_observables != observables, axis=1))

print(f"{num_mistakes}/{num_shots}")

Sinter integration

To integrate with sinter, you can use the ilp_decoder.ILPSinterDecoder class, which inherits from sinter.Decoder. To use it, you can use the custom_decoders argument when using sinter.collect:

import sinter
from ilp_decoder import ILPSinterDecoder, sinter_decoders

samples = sinter.collect(
    num_workers=4,
    max_shots=1_000_000,
    max_errors=1000,
    tasks=generate_example_tasks(),
    decoders=['ilp'],
    custom_decoders=sinter_decoders()
    print_progress=True,
)

Advanced ILP solvers

You can use whatever ILP solver supported by cvxpy instead of the default one highs, though highs should be one of the fastest open-source solver available. If you have access to gurobi, you may gain some extra speed boost (not tested).

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

ilp_decoder-0.2.0.tar.gz (71.9 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

ilp_decoder-0.2.0-py3-none-any.whl (12.7 kB view details)

Uploaded Python 3

File details

Details for the file ilp_decoder-0.2.0.tar.gz.

File metadata

  • Download URL: ilp_decoder-0.2.0.tar.gz
  • Upload date:
  • Size: 71.9 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: uv/0.6.10

File hashes

Hashes for ilp_decoder-0.2.0.tar.gz
Algorithm Hash digest
SHA256 b4bdce932391f1302a6d8ddea6ec37856183a294052421d65433e5db44c2bac4
MD5 992083103dbebeb7edeb7478ce5a759e
BLAKE2b-256 e8f8f2e69b7672c5670b617b3b9ffe8b58a6702dc878d42be7c73fe9582914ca

See more details on using hashes here.

File details

Details for the file ilp_decoder-0.2.0-py3-none-any.whl.

File metadata

File hashes

Hashes for ilp_decoder-0.2.0-py3-none-any.whl
Algorithm Hash digest
SHA256 b9ecdc3351c12c51b1f35a424951d313fa3c8d4786e48110f8ebcfd73a4c9a06
MD5 69d577bbaf697178a7aa65ec8d049cfd
BLAKE2b-256 4c4a8925545f79b5bbafb53b62e1fcaa64aa11c11f690b41aaa896f2ffae935b

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page