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
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file ilp_decoder-0.2.2.tar.gz.
File metadata
- Download URL: ilp_decoder-0.2.2.tar.gz
- Upload date:
- Size: 124.3 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: uv/0.8.14
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
86706d01ea94ca3c609f1815d491852e5b9c7827f3a0c80b8f64d07152e7c3ef
|
|
| MD5 |
0f63ae9ead14c933594622b35d47e409
|
|
| BLAKE2b-256 |
e0b4ab9b0a47955d16212cd5b8f9e7eedb454110fe895b6d2e77056a23d15744
|
File details
Details for the file ilp_decoder-0.2.2-py3-none-any.whl.
File metadata
- Download URL: ilp_decoder-0.2.2-py3-none-any.whl
- Upload date:
- Size: 12.8 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: uv/0.8.14
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
c992ec8b5536e70961201e68be111b39ed9d27609f4f58d95eb01039749784dd
|
|
| MD5 |
109551b2c9a585b45b156fa79a5a372d
|
|
| BLAKE2b-256 |
a665ac5c1b7d830de93d69d061ffaa5b9ae7d011501b09fb7637a86013326286
|