Larq
Larq: An Open-Source Library for Training Binarized Neural Networks - Published in JOSS (2020)
Science Score: 95.0%
This score indicates how likely this project is to be science-related based on various indicators:
-
○CITATION.cff file
-
✓codemeta.json file
Found codemeta.json file -
✓.zenodo.json file
Found .zenodo.json file -
✓DOI references
Found 4 DOI reference(s) in README and JOSS metadata -
✓Academic publication links
Links to: joss.theoj.org -
✓Committers with academic emails
2 of 18 committers (11.1%) from academic institutions -
○Institutional organization owner
-
✓JOSS paper metadata
Published in Journal of Open Source Software
Keywords
Keywords from Contributors
Scientific Fields
Repository
An Open-Source Library for Training Binarized Neural Networks
Basic Info
- Host: GitHub
- Owner: larq
- License: apache-2.0
- Language: Python
- Default Branch: main
- Homepage: https://larq.dev
- Size: 1020 KB
Statistics
- Stars: 720
- Watchers: 29
- Forks: 85
- Open Issues: 30
- Releases: 38
Topics
Metadata Files
README.md
Larq is an open-source deep learning library for training neural networks with extremely low precision weights and activations, such as Binarized Neural Networks (BNNs).
Existing deep neural networks use 32 bits, 16 bits or 8 bits to encode each weight and activation, making them large, slow and power-hungry.
This prohibits many applications in resource-constrained environments. Larq is the first step towards solving this. It is designed to provide an easy to use, composable way to train BNNs (1 bit) and other types of Quantized Neural Networks (QNNs) and is based on the tf.keras interface. Note that efficient inference using a trained BNN requires the use of an optimized inference engine; we provide these for several platforms in Larq Compute Engine.
Larq is part of a family of libraries for BNN development; you can also check out Larq Zoo for pretrained models and Larq Compute Engine for deployment on mobile and edge devices.
Getting Started
To build a QNN, Larq introduces the concept of quantized layers and quantizers. A quantizer defines the way of transforming a full precision input to a quantized output and the pseudo-gradient method used for the backwards pass. Each quantized layer requires an input_quantizer and a kernel_quantizer that describe the way of quantizing the incoming activations and weights of the layer respectively. If both input_quantizer and kernel_quantizer are None the layer is equivalent to a full precision layer.
You can define a simple binarized fully-connected Keras model using the Straight-Through Estimator the following way:
python
model = tf.keras.models.Sequential(
[
tf.keras.layers.Flatten(),
larq.layers.QuantDense(
512, kernel_quantizer="ste_sign", kernel_constraint="weight_clip"
),
larq.layers.QuantDense(
10,
input_quantizer="ste_sign",
kernel_quantizer="ste_sign",
kernel_constraint="weight_clip",
activation="softmax",
),
]
)
This layer can be used inside a Keras model or with a custom training loop.
Examples
Check out our examples on how to train a Binarized Neural Network in just a few lines of code:
Installation
Before installing Larq, please install:
- Python version
3.7,3.8,3.9, or3.10 - Tensorflow version
1.14,1.15,2.0,2.1,2.2,2.3,2.4,2.5,2.6,2.7,2.8,2.9, or2.10:shell pip install tensorflow # or tensorflow-gpu
You can install Larq with Python's pip package manager:
shell
pip install larq
About
Larq is being developed by a team of deep learning researchers and engineers at Plumerai to help accelerate both our own research and the general adoption of Binarized Neural Networks.
Owner
- Name: larq
- Login: larq
- Kind: organization
- Location: London - Amsterdam
- Website: https://larq.dev
- Repositories: 7
- Profile: https://github.com/larq
An Open-Source Deep Learning Library for Training Binarized Neural Networks
JOSS Publication
Larq: An Open-Source Library for Training Binarized Neural Networks
Authors
Plumerai Research
Plumerai Research
Editor
Yuan Tang
Tags
python tensorflow keras deep-learning machine-learning binarized-neural-networks quantized-neural-networks efficient-deep-learningPapers & Mentions
Total mentions: 1
FPGA-Based Acceleration on Additive Manufacturing Defects Inspection
- DOI: 10.3390/s21062123
- OpenAlex ID: https://openalex.org/W3137707937
- Published: March 2021
GitHub Events
Total
- Issues event: 9
- Watch event: 21
- Issue comment event: 27
- Fork event: 2
Last Year
- Issues event: 9
- Watch event: 21
- Issue comment event: 27
- Fork event: 2
Committers
Last synced: 5 months ago
Top Committers
| Name | Commits | |
|---|---|---|
| Lukas Geiger | l****r | 362 |
| dependabot[bot] | 4****] | 212 |
| dependabot-preview[bot] | 2****] | 72 |
| Leon Overweel | l****l@g****m | 20 |
| Adam Hillier | 7****r | 14 |
| Koen Helwegen | k****n@p****m | 12 |
| MariaHeuss | 5****s | 10 |
| Jelmer Neeven | 1****n | 8 |
| James Widdicombe | j****6@g****m | 8 |
| Koen Helwegen | k****n@g****m | 6 |
| Arash Bakhtiari | b****r@i****e | 3 |
| Simon Brugman | s****n | 2 |
| Peter Fackeldey | p****y@r****e | 1 |
| Joschua-Conrad | 3****d | 1 |
| Tim de Bruin | t****m@p****m | 1 |
| Tom Bannink | t****k@g****m | 1 |
| imgbot[bot] | 3****] | 1 |
| rnusselder | 4****r | 1 |
Committer Domains (Top 20 + Academic)
Issues and Pull Requests
Last synced: 4 months ago
All Time
- Total issues: 45
- Total pull requests: 157
- Average time to close issues: 22 days
- Average time to close pull requests: 10 days
- Total issue authors: 11
- Total pull request authors: 6
- Average comments per issue: 1.36
- Average comments per pull request: 0.32
- Merged pull requests: 130
- Bot issues: 0
- Bot pull requests: 73
Past Year
- Issues: 6
- Pull requests: 0
- Average time to close issues: 6 days
- Average time to close pull requests: N/A
- Issue authors: 3
- Pull request authors: 0
- Average comments per issue: 5.0
- Average comments per pull request: 0
- Merged pull requests: 0
- Bot issues: 0
- Bot pull requests: 0
Top Authors
Issue Authors
- lgeiger (28)
- koenhelwegen (5)
- Oslomayor (2)
- x7779 (2)
- Sameeksha0709 (2)
- jamescook106 (1)
- vfdev-5 (1)
- 17387990341 (1)
- dependabot[bot] (1)
- Junshuai1024 (1)
- Nannigalaxy (1)
Pull Request Authors
- dependabot[bot] (97)
- lgeiger (74)
- jamescook106 (4)
- koenhelwegen (3)
- RanHomri (1)
- yotamazriel (1)
Top Labels
Issue Labels
Pull Request Labels
Packages
- Total packages: 1
-
Total downloads:
- pypi 1,315 last-month
- Total dependent packages: 1
- Total dependent repositories: 8
- Total versions: 37
- Total maintainers: 1
pypi.org: larq
An Open Source Machine Learning Library for Training Binarized Neural Networks
- Homepage: https://larq.dev/
- Documentation: https://larq.readthedocs.io/
- License: Apache 2.0
-
Latest release: 0.13.3
published over 2 years ago
Rankings
Maintainers (1)
Dependencies
- dataclasses *
- importlib-metadata *
- numpy *
- terminaltables >=3.1.0
- actions/checkout v3 composite
- actions/setup-python v4 composite
- actions/checkout v3 composite
- actions/setup-python v4 composite
- pypa/gh-action-pypi-publish master composite
- toolmantim/release-drafter v5.22.0 composite
- actions/checkout v3 composite
- actions/setup-python v4 composite