Releases | Open Source Science

Larq - v0.13.3

:bug: Bug Fixes

Add QuantizedVariableSpec to QuantizedVariable (#838) @lgeiger

:arrow_up: Dependencies

Bump black from 23.3.0 to 23.7.0 (#839) @dependabot

Scientific Software - Peer-reviewed - Python
Published by github-actions[bot] over 2 years ago

Larq - v0.13.2

:tada: Features

Support TensorFlow 2.13 (#831) @lgeiger

:arrow_up: Dependencies

Bump toolmantim/release-drafter from 5.23.0 to 5.24.0 (#834) @dependabot
Update pytest requirement from ==7.3.* to >=7.3,<7.5 (#833) @dependabot
Update pytest-mock requirement from ==3.10.* to >=3.10,<3.12 (#832) @dependabot

Scientific Software - Peer-reviewed - Python
Published by github-actions[bot] over 2 years ago

Larq - v0.13.1

:tada: Features

Support TensorFlow 2.12 (#829) @lgeiger

:constructionworkerman: Internal Improvements

Update TensorFlow dependencies on CI (#826) @lgeiger

:arrow_up: Dependencies

Update pytest-cov requirement from ==4.0.* to >=4.0,<4.2 (#828) @dependabot
Update pytest requirement from ==7.2.* to >=7.2,<7.4 (#825) @dependabot
Update pytest-xdist requirement from ==3.0.* to >=3.0,<3.3 (#822) @dependabot
Bump toolmantim/release-drafter from 5.21.1 to 5.23.0 (#819, #823) @dependabot
Update isort requirement from ==5.10.* to >=5.10,<5.12 (#818) @dependabot
Bump black from 22.10.0 to 23.3.0 (#817, #821, #824) @dependabot
Update flake8 requirement from ==5.0.* to >=5.0,<6.1 (#814) @dependabot

Scientific Software - Peer-reviewed - Python
Published by github-actions[bot] over 2 years ago

Larq - v0.13.0

:warning: Breaking Changes :warning:

Drop support for Python 3.6 (#811) @lgeiger

:tada: Features

Support TensorFlow 2.11 (#810) @lgeiger

:constructionworkerman: Internal Improvements

Pin test requirements (#812) @lgeiger
Update TensorFlow on CI (#795) @lgeiger
Test against TensorFlow 2.10 (#790) @lgeiger
:arrow_up: TF to fix CI (#775) @lgeiger
Add TensorFlow 2.9.0 to CI (#770) @lgeiger
Test against Python 3.10 (#765) @lgeiger
Test against TF 2.9.0rc0 (#762) @lgeiger
Test against TensorFlow 2.8 (#747) @lgeiger

:arrow_up: Dependencies

Bump pytype from 2021.10.25 to 2022.10.26 (#809, #805, #800, #799, #796, #793, #791, #789, #787, #785, #783, #780, #779, #773, #771, #769) @dependabot
Update pytest-xdist requirement from <2.5,>=1.30 to >=1.30,<3.1 (#808, #738) @dependabot
Update pytest requirement from <6.3.0,>=5.2.4 to >=5.2.4,<7.3.0 (#807, #758, #751) @dependabot
Update pytest-mock requirement from <3.10,>=2.0 to >=2.0,<3.11 (#803, #801, #781, #745) @dependabot
Update pytest-cov requirement from <3.1.0,>=2.8.1 to >=2.8.1,<4.1.0 (#802) @dependabot
Bump toolmantim/release-drafter from 5.15.0 to 5.21.1 (#806, #788, #768, #757, #752, #749, #744, #743, #740) @dependabot
Bump black from 21.11b1 to 22.10.0 (#804, #794, #782, #759, #746, #737) @dependabot
Update flake8 requirement from <4.1.0,>=3.7.9 to >=3.7.9,<5.1.0 (#786) @dependabot
Bump actions/setup-python from 2.3.1 to 4 (#776, #755, #750) @dependabot
Bump actions/checkout from 2 to 3 (#756) @dependabot

Scientific Software - Peer-reviewed - Python
Published by github-actions[bot] about 3 years ago

Larq - v0.12.2

:bug: Bug Fixes

Do not map variables according to names in CaseOptimizer (#735) @lgeiger
Prevent CaseOptimizer from creating unused vars (#734) @lgeiger
Support nested models in ModelProfile (#718) @jneeven

:constructionworkerman: Internal Improvements

:arrow_up: TensorFlow on CI (#725) @lgeiger
Add TensorFlow 2.7.0rc0 to CI matrix (#710) @lgeiger

:arrow_up: Dependencies

Bump actions/setup-python from 2.2.2 to 2.3.1 (#728, #733) @dependabot
Bump black from 21.9b0 to 21.11b1 (#721, #729, #730) @dependabot
Bump isort from 5.9.3 to 5.10.1 (#724, #727) @dependabot
Bump pytype from 2021.10.4 to 2021.10.25 (#712, #715, #717) @dependabot
Update flake8 requirement from <3.10.0,>=3.7.9 to >=3.7.9,<4.1.0 (#711) @dependabot

Scientific Software - Peer-reviewed - Python
Published by github-actions[bot] about 4 years ago

Larq - v0.12.1

:bug: Bug Fixes

Correctly pass output_padding to transposed convolutions (#707) @lgeiger

:book: Documentation

Properly pass groups and dilation_rate parameters to convolutional layers (#708) @lgeiger

:constructionworkerman: Internal Improvements

Properly pass groups and dilation_rate parameters to convolutional layers (#708) @lgeiger
Rename master -> main (#701) @lgeiger
Update CI to TensorFlow 2.6.0 and Python 3.9 (#691) @lgeiger
Add TF 2.6.0rc0 to CI (#682) @lgeiger
Update CI to TF 2.5.0 stable (#674) @lgeiger

:arrow_up: Dependencies

Update pytest-cov requirement from <2.12.0,>=2.8.1 to >=2.8.1,<3.1.0 (#667, #705) @dependabot
Bump pytype from 2021.5.4 to 2021.10.4 (#663, #665, #668, #670, #671, #680, #687, #688, #690, #692, #700, #706) @dependabot
Update pytest-xdist requirement from <2.3,>=1.30 to >=1.30,<2.5 (#679, #703) @dependabot
Bump black from 21.5b0 to 21.9b0 (#686, #694, #702) @dependabot
Bump isort from 5.8.0 to 5.9.3 (#664, #675, #678, #681, #684, #689) @dependabot
Update packaging requirement from <21.0,>=19.2 to >=19.2,<22.0 (#683) @dependabot

Scientific Software - Peer-reviewed - Python
Published by github-actions[bot] over 4 years ago

Larq - v0.12.0

:tada: Features

Add DoReFa Weight Quantizer (#651) @Joschua-Conrad

:book: Documentation

Document support for TF 2.5 and Python 3.9 (#661) @lgeiger
Add DoReFa plot for mode="weights" (#660) @lgeiger
Remove outdated badges (#638) @lgeiger

:constructionworkerman: Internal Improvements

Add TensorFlow 2.5.0rc0 and Python 3.9 to CI matrix (#647) @lgeiger
[callbacks.py] switch to python3 super() call (#646) @pfackeldey

:arrow_up: Dependencies

Bump black from 20.8b1 to 21.5b0 (#653, #656, #657, #659) @dependabot
Bump pytype from 2020.10.8 to 2021.5.4 (#632, #634, #635, #643, #644, #648, #650, #655, #658) @dependabot
Update pytest-mock requirement from <3.6,>=2.0 to >=2.0,<3.7 (#654) @dependabot
Bump actions/setup-python from v2.2.1 to v2.2.2 (#649) @dependabot
Bump isort from 5.7.0 to 5.8.0 (#642) @dependabot
Bump toolmantim/release-drafter from v5.14.0 to v5.15.0 (#639) @dependabot
Update flake8 requirement from <3.9.0,>=3.7.9 to >=3.7.9,<3.10.0 (#636) @dependabot

Scientific Software - Peer-reviewed - Python
Published by github-actions[bot] over 4 years ago

Larq - v0.11.2

:bug: Bug Fixes

Initialise CaseOptimizer.iterations which is required in TF 2.3+ (#629) @lgeiger

Scientific Software - Peer-reviewed - Python
Published by github-actions[bot] almost 5 years ago

Larq - v0.11.1

:bug: Bug Fixes

Rename Quantizer type to prevent conflict with base class (#627) @lgeiger

:constructionworkerman: Internal Improvements

Favour lq.quantizers.NoOp over NoOpQuantizer (#626) @lgeiger

Scientific Software - Peer-reviewed - Python
Published by github-actions[bot] almost 5 years ago

Larq - v0.11.0

:warning: Breaking Changes :warning:

Add public base class for defining quantizers and hide BaseQuantizer (#623) @lgeiger

:bug: Bug Fixes

Call super().build() in Quantizers.build() (#621) @lgeiger

:constructionworkerman: Internal Improvements

Build flip_ratio metric in Layer.build (#622) @lgeiger

:arrow_up: Dependencies

Bump toolmantim/release-drafter from v5.13.0 to v5.14.0 (#624) @dependabot

Scientific Software - Peer-reviewed - Python
Published by github-actions[bot] almost 5 years ago

Larq - v0.10.3

:bug: Bug Fixes

Support 1d ops in model profile (#614) @jneeven

:book: Documentation

Add TF2.4 to the README. (#617) @AdamHillier

:constructionworkerman: Internal Improvements

Update TF versions in unit tests. (#616) @AdamHillier
Add TF 2.4 stable to CI (#602) @lgeiger
Relax importlib-metadata dependency constraint for older Python < 3.8 (#599) @lgeiger
Add TensorFlow 2.4.0rc0 to CI matrix (#598) @lgeiger

:arrow_up: Dependencies

Update pytest-cov requirement from <2.11.0,>=2.8.1 to >=2.8.1,<2.12.0 (#612) @dependabot
Update pytest-mock requirement from <3.5,>=2.0 to >=2.0,<3.6 (#609) @dependabot
Bump isort from 5.6.4 to 5.7.0 (#608) @dependabot
Bump toolmantim/release-drafter from v5.12.1 to v5.13.0 (#606) @dependabot
Bump actions/setup-python from v2.2.0 to v2.2.1 (#605) @dependabot
Bump actions/setup-python from v2.1.4 to v2.2.0 (#604) @dependabot
Update pytest-mock requirement from <3.4,>=2.0 to >=2.0,<3.5 (#603) @dependabot
Update pytest-xdist requirement from <2.2,>=1.30 to >=1.30,<2.3 (#601) @dependabot
Update pytest requirement from <6.2.0,>=5.2.4 to >=5.2.4,<6.3.0 (#600) @dependabot

Scientific Software - Peer-reviewed - Python
Published by github-actions[bot] almost 5 years ago

Larq - v0.10.2

:constructionworkerman: Internal Improvements

Add __version__ property. Closes #595. (#596) @AdamHillier
Upgrade TensorFlow versions on CI (#580) @lgeiger

:arrow_up: Dependencies

Bump toolmantim/release-drafter from v5.12.0 to v5.12.1 (#594) @dependabot
Bump toolmantim/release-drafter from v5.11.0 to v5.12.0 (#593) @dependabot
Bump actions/setup-python from v2.1.3 to v2.1.4 (#592) @dependabot
Bump isort from 5.6.3 to 5.6.4 (#591) @dependabot
Bump isort from 5.6.1 to 5.6.3 (#590) @dependabot
Bump pytype from 2020.9.29 to 2020.10.8 (#588) @dependabot
Bump isort from 5.6.0 to 5.6.1 (#589) @dependabot
Bump isort from 5.5.4 to 5.6.0 (#587) @dependabot
Bump actions/setup-python from v2.1.2 to v2.1.3 (#585) @dependabot
Bump pytype from 2020.9.24 to 2020.9.29 (#583) @dependabot
Bump isort from 5.5.3 to 5.5.4 (#584) @dependabot
Update snapshottest requirement from ~=0.5.1 to >=0.5.1,<0.7.0 (#582) @dependabot
Update pytest requirement from <6.1.0,>=5.2.4 to >=5.2.4,<6.2.0 (#581) @dependabot

Scientific Software - Peer-reviewed - Python
Published by github-actions[bot] about 5 years ago

Larq -

:bug: Bug Fixes

Fix variable saving on tf-nightly (#578) @lgeiger
Raise exception when causal padding is used with pad_values != 0 (#575) @lgeiger
Improve model summary for TF layers and multi output layers (#572) @lgeiger
Fix distributed quantized variable assignment in TF 2.x (#556) @lgeiger

:book: Documentation

Fix indentation of DoReFa warning (#576) @lgeiger
Emphasize importance of LCE in README.md (#564) @leonoverweel

:constructionworkerman: Internal Improvements

Only warn for missing kernel constraints when quantizers are binary (#577) @lgeiger
Run linting using Python 3.8 on CI (#557) @lgeiger

:arrow_up: Dependencies

Bump pytype from 2020.9.16 to 2020.9.24 (#573) @dependabot
Bump isort from 5.5.2 to 5.5.3 (#570) @dependabot
Bump pytype from 2020.9.14 to 2020.9.16 (#569) @dependabot
Bump pytype from 2020.8.28 to 2020.9.14 (#568) @dependabot
Bump isort from 5.5.1 to 5.5.2 (#567) @dependabot
Bump isort from 5.5.0 to 5.5.1 (#566) @dependabot
Bump isort from 5.4.2 to 5.5.0 (#565) @dependabot
Bump pytype from 2020.8.17 to 2020.8.28 (#563) @dependabot
Bump black from 19.10b0 to 20.8b1 (#561) @dependabot
Update pytest-xdist requirement from <2.1,>=1.30 to >=1.30,<2.2 (#560) @dependabot
Update pytest-mock requirement from <3.3,>=2.0 to >=2.0,<3.4 (#559) @dependabot
Bump pytype from 2020.8.10 to 2020.8.17 (#558) @dependabot
Update pytest-xdist requirement from <1.35,>=1.30 to >=1.30,<2.1 (#555) @dependabot
Bump isort from 5.4.1 to 5.4.2 (#554) @dependabot
Bump isort from 5.4.0 to 5.4.1 (#552) @dependabot
Bump isort from 5.3.2 to 5.4.0 (#551) @dependabot
Bump actions/setup-python from v2.1.1 to v2.1.2 (#550) @dependabot
Bump pytype from 2020.7.30 to 2020.8.10 (#549) @dependabot
Bump isort from 5.3.0 to 5.3.2 (#548) @dependabot
Bump isort from 5.2.2 to 5.3.0 (#547) @dependabot
Bump isort from 5.2.1 to 5.2.2 (#545) @dependabot
Bump pytype from 2020.7.24 to 2020.7.30 (#546) @dependabot
Bump isort from 5.2.0 to 5.2.1 (#543) @dependabot
Update pytest requirement from <5.5.0,>=5.2.4 to >=5.2.4,<6.1.0 (#542) @dependabot
Run test against TensorFlow 2.3 stable on CI (#540) @lgeiger
Update pytest-xdist requirement from <1.34,>=1.30 to >=1.30,<1.35 (#541) @dependabot
Bump isort from 5.1.4 to 5.2.0 (#539) @dependabot
Bump pytype from 2020.7.20 to 2020.7.24 (#538) @dependabot
Bump actions/setup-python from v2.1.0 to v2.1.1 (#536) @dependabot
Bump pytype from 2020.7.14 to 2020.7.20 (#537) @dependabot
Bump isort from 5.1.1 to 5.1.4 (#535) @dependabot
Bump actions/setup-python from v2.0.2 to v2.1.0 (#534) @dependabot
Bump actions/setup-python from v2.0.1 to v2.0.2 (#533) @dependabot
Bump isort from 5.1.0 to 5.1.1 (#532) @dependabot
Bump pytype from 2020.6.26 to 2020.7.14 (#531) @dependabot
Bump isort from 5.0.9 to 5.1.0 (#530) @dependabot
Update pytest-mock requirement from <3.2,>=2.0 to >=2.0,<3.3 (#528) @dependabot
Bump isort from 5.0.7 to 5.0.9 (#529) @dependabot
Run unittests against TF 2.3rc1 (#527) @lgeiger
Update pytest-xdist requirement from <1.33,>=1.30 to >=1.30,<1.34 (#525) @dependabot
Bump isort from 5.0.6 to 5.0.7 (#526) @dependabot
Bump isort from 5.0.4 to 5.0.6 (#524) @dependabot

Scientific Software - Peer-reviewed - Python
Published by github-actions[bot] over 5 years ago

Larq - v0.10.0

:tada: Features

Add support for TensorFlow 2.3 (#498) @lgeiger

:bug: Bug Fixes

Fix saving in SavedModel format (#514) @lgeiger
Fix QuantizedVariable.set_shape and improve method delegation test (#518) @lgeiger
Fix metric typo (#497) @koenhelwegen

:book: Documentation

Add TensorFlow 2.3 to list of supported versions (#520) @lgeiger
Improve alignment of LaTeX formulas (#501) @lgeiger
Apply Python Black formatting to docstring examples (#500) @leonoverweel
Add warning to DoReFa docstring (#494) @jneeven

:constructionworkerman: Internal Improvements

Add test for set_precision decorator (#517) @lgeiger
Upgrade isort (#516) @lgeiger
Always upload coverage on CI to reduce flaky code coverage reports (#515) @lgeiger
Run distribution strategy tests with multiple virtual CPUs (#507) @lgeiger
Cleanup tests (#506) @lgeiger

:arrow_up: Dependencies

Bump isort from 5.0.1 to 5.0.4 (#519, #521) @dependabot
Upgrade pytype to autoselect jobs on CI (#513) @lgeiger
Bump toolmantim/release-drafter from v5.7.0 to v5.11.0 (#512) @dependabot
Bump actions/setup-python from v1 to v2.0.1 (#511) @dependabot
Create Dependabot config file (#510) @dependabot-preview
Update pytype requirement from <2020.3.0,>=2020.01.07 to >=2020.01.07,<2020.7.0 (#499) @dependabot-preview
Update pytest-cov requirement from ~=2.8.1 to >=2.8.1,<2.11.0 (#495, #505) @dependabot-preview

Scientific Software - Peer-reviewed - Python
Published by github-actions[bot] over 5 years ago

Larq - v0.9.6

:bug: Bug Fixes

Don't lazily compute isnative_padding (#492) @lgeiger
Correctly compute one-padding shape to prevent recreating conv_op (#491) @lgeiger

:book: Documentation

Fix references for QuantConv3DTranspose (#490) @leonoverweel

:arrow_up: Dependencies

Update flake8 requirement from ~=3.7.9 to >=3.7.9,<3.9.0 (#489) @dependabot-preview

Scientific Software - Peer-reviewed - Python
Published by github-actions[bot] over 5 years ago

Larq -

:bug: Bug Fixes

Fix case optimizer when used with distributed mixed precision loss scaling (#485) @lgeiger
Fix larq summary for subclassed models (#480) @lgeiger

:book: Documentation

Reformat docs for keras-autodoc (#482) @leonoverweel
Add TF 2.2.0 to CI test matrix (#486) @lgeiger

:arrow_up: Dependencies

Update pytest-xdist requirement from <1.32,>=1.30 to >=1.30,<1.33 (#481) @dependabot-preview
Update pytest-mock requirement from <3.1,>=2.0 to >=2.0,<3.2 (#476) @dependabot-preview

Scientific Software - Peer-reviewed - Python
Published by github-actions[bot] over 5 years ago

Larq -

:tada: Features

Add support for experimental_aggregate_gradients (#466) @lgeiger

:book: Documentation

Replace arXiv URLs with https version (#474) @lgeiger
Fix typo in metrics (#473) @lgeiger

:constructionworkerman: Internal Improvements

Upload coverage without token (#472) @lgeiger
Implement compute_output_shape in base quantizer (#471) @lgeiger
Add Python 3.8 to CI matrix (#469) @lgeiger
Make quantiser names consistent. Closes #467. (#468) @AdamHillier

:arrow_up: Dependencies

Update pytest-mock requirement from ~=2.0.0 to >=2.0,<3.1 (#470) @dependabot-preview

Scientific Software - Peer-reviewed - Python
Published by github-actions[bot] over 5 years ago

Larq -

:bug: Bug Fixes

Use static tensor shapes if possible for one-padding (#463) @lgeiger

Scientific Software - Peer-reviewed - Python
Published by github-actions[bot] almost 6 years ago

Larq - v0.9.2

:bug: Bug Fixes

Support unknown input shapes with one padding (#461) @lgeiger

:arrow_up: Dependencies

Update pytest requirement from <5.4.0,>=5.2.4 to >=5.2.4,<5.5.0 (#460) @dependabot-preview

Scientific Software - Peer-reviewed - Python
Published by github-actions[bot] almost 6 years ago

Larq -

:bug: Bug Fixes

Fix deserialization of quantizer functions (#458) @lgeiger

Scientific Software - Peer-reviewed - Python
Published by github-actions[bot] almost 6 years ago

Larq -

:warning: Breaking Changes :warning:

Consolidate context managers under lq.context (#450) @lgeiger
- lq.metrics.scope was renamed to lq.context.metrics_scope
- lq.context.quantized_scope can now be used to retrieve the quantized result from variables
Add NoOpQuantizer for use with custom optimizer (#442) @lgeiger
- Bop won't automatically collect variables from quantized layers anymore.
- NoOpQuantizer can be used to explicitly mark variables to be targeted by Bop: python no_op_quantizer = lq.quantizers.NoOpQuantizer(precision=1) layer = lq.layers.QuantDense(16, kernel_quantizer=no_op_quantizer)
Move custom metrics from layers to quantizers (#402) @leonoverweel
- Layers don't have a metrics metrics keyword anymore, it is now part of the quantizers.
- For the reasoning behind this change read RFC #343
Add Quantized Variables to handle quantization (#312, #359) @lgeiger
- For more information see RFC #306
All builtin quantizers now subclass from keras.layers.Layer (#387) @lgeiger
- For more information see RFC #343
Change default hyperparameters of Bop (#363) @MariaHeuss

:tada: Features

Add support for non-zero padding (#438) @lgeiger
Correctly calculate Batchnorm in profile (#454) @koenhelwegen
Add method for printing human-readable memory sizes (#439) @leonoverweel
Model size for 8-bit quantized full-precision weights (#449) @leonoverweel
Make HyperparameterScheduler compatible with CaseOptimizer (#356) @MariaHeuss

:book: Documentation

Fix ordering of docstring arguments scheduler (#455) @koenhelwegen
Add Larq meta info to README.md (#445) @leonoverweel
Fix threshold description in (#435) @timdebruin
Move docs to larq/docs (#433) @lgeiger
Align the navbar icons. (#430) @AdamHillier
Remove outdated reference to tutorial from README, closes #431 (#432) @AdamHillier
Workaround #422 (#424) @lgeiger
Fix more links (#423) @lgeiger
Restructure docs and add compute-engine tab (#418) @lgeiger
Fix weight_clip documentation issue (#371) @lgeiger
Remove outdated line in Bop documentation (#367) @Tombana
Fix link to Bop in docs. (#380) @AdamHillier
Fix typo in case optimizer docs (#382) @lgeiger
Fix typo in DoReFa docs. (#392) @AdamHillier

:constructionworkerman: Internal Improvements

Case opt check (#446) @leonoverweel
Remove unnecessary # from format string (#447) @leonoverweel
Move memory string formatting into utils (#444) @leonoverweel
Simplify quantizer metrics (#437) @lgeiger
Cleanup base layers (#436) @lgeiger
Lazily build metrics (#404) @leonoverweel
Simplify assign operations in Bop (#357, #306) @lgeiger
Test against TF 2.1.0 (#378) @lgeiger
(Partially) add type hints. See #309. (#391) @AdamHillier
Remove gast from requirements.txt (#395) @lgeiger
Lazily build metrics (#404) @leonoverweel
Add quantized variable fixes from TF upstream (#452) @lgeiger
Add CI job to test larq against TensorFlow 2.2rc0 (#451) @lgeiger
Simplify release process (#456) @lgeiger

:arrow_up: Dependencies

Upgrade github actions (#453) @lgeiger
Bump mknotebooks to 0.1.16 (#429) @dependabot-preview
Bump mkdocs-material from 4.6.2 to 4.6.3 (#416) @dependabot-preview
Bump mkdocs-minify-plugin from 0.2.1 to 0.2.3 (#415) @dependabot-preview
Bump mkdocs-material to 4.6.3 (#413) @dependabot-preview
Update pytype requirement from <2020.2.0,>=2020.01.07 to >=2020.01.07,<2020.3.0 (#408) @dependabot-preview
Bump pandas to 1.0.1 (#407) @dependabot-preview
Bump matplotlib from 3.1.2 to 3.1.3 (#403) @dependabot-preview

Scientific Software - Peer-reviewed - Python
Published by github-actions[bot] almost 6 years ago

Larq -

:tada: Features

Support per-step scheduling in HyperParameterScheduler (#385) @jneeven

:bug: Bug Fixes

Fix optimizer reloading of CaseOptimizer with tf.train.Checkpoint (#398) @lgeiger

Scientific Software - Peer-reviewed - Python
Published by AdamHillier almost 6 years ago

Larq -

:book: Documentation

Fix paper.bib (#384) @lgeiger
Add missing DOIs to paper references (#383) @lgeiger
Address paper review comments (#370) @lgeiger
Cite NeurIPS paper instead of arXiv version (#355) @lgeiger
Cite conference versions of the papers (#352) @lgeiger
Fix paper typo (#351) @lgeiger
Fix typos in paper (#345) @lgeiger

Scientific Software - Peer-reviewed - Python
Published by lgeiger almost 6 years ago

Larq -

:bug: Bug Fixes

Fix reloading of case optimizer weights (#340) @lgeiger

:book: Documentation

Second paper draft (#329) @lgeiger

:constructionworkerman: Internal Improvements

Test against TF 2.1rc0 (#339) @lgeiger

:arrow_up: Dependencies

Update pytype requirement from ~=2019.10.17 to >=2019.10.17,<2019.12.0 (#338) @dependabot-preview
Bump altair from 3.2.0 to 3.3.0 (#337) @dependabot-preview

Scientific Software - Peer-reviewed - Python
Published by github-actions[bot] about 6 years ago

Larq -

:bug: Bug Fixes

Fix reloading of Bop when used with schedules (#334) @lgeiger
Fix try except in model summary (#332) @lgeiger

:book: Documentation

Fix examples in documentation for CaseOptimizer and Bop (#331) @leonoverweel

:constructionworkerman: Internal Improvements

Improve linting (#328) @AdamHillier

:arrow_up: Dependencies

Bump matplotlib from 3.1.1 to 3.1.2 (#333) @dependabot-preview
Update pytest requirement from ~=5.2.4 to >=5.2.4,<5.4.0 (#330) @dependabot-preview
Bump mkdocs-material from 4.4.3 to 4.5.0 (#327) @dependabot-preview
Bump pymdown-extensions from 6.1 to 6.2 (#326) @dependabot-preview
Bump pytest from 5.2.3 to 5.2.4 (#325) @dependabot-preview

Scientific Software - Peer-reviewed - Python
Published by github-actions[bot] about 6 years ago

Larq -

:warning: Breaking Changes :warning:

Drop TensorFlow 1.13 (#317) @leonoverweel
Refactor Bop into Bop + CaseOptimizer (#304) @leonoverweel

:bug: Bug Fixes

Fix zero division bug in summary for weightless models (#302) @jneeven

:book: Documentation

Add link to NeurIPS poster (#318) @lgeiger
Enhancements for CONTRIBUTING.md shell commands + adding editors to .gitignore (#303) @leonoverweel
Add google analytics tracking (#295) @lgeiger
Use animated logo in the README (#291) @lgeiger
Update Release Docs (#290) @lgeiger
Fix links on front page (#288) @lgeiger

:constructionworkerman: Internal Improvements

Quantizer layer tests (#322) @leonoverweel
Group and organize quantizer tests (#321) @leonoverweel
Simplify layer tests (#320) @lgeiger
Refactor tests to not use TensorFlow TestCase (#319) @leonoverweel
Fix log count check in layer warnings tests (#316) @leonoverweel
Add pytest fixture to run tests in eager and graph mode (#313) @lgeiger
Run unit tests in parallel using pytest-xdist (#308) @lgeiger
Test against TensorFlow 1.15.0 stable (#294) @lgeiger
Heaviside Function (#293) @jamescook106
Cleanup lq.math tests (#292) @jamescook106

:arrow_up: Dependencies

Bump pytest from 5.2.2 to 5.2.3 (#323) @dependabot-preview
Bump black (#310) @lgeiger
Upgrade MathJax@3 (#307) @lgeiger
Bump pandas from 0.25.2 to 0.25.3 (#305) @dependabot-preview
Bump pytest from 5.2.1 to 5.2.2 (#300) @dependabot-preview
Bump pillow from 6.2.0 to 6.2.1 (#298) @dependabot-preview
Bump pandas from 0.25.1 to 0.25.2 (#296) @dependabot-preview

Scientific Software - Peer-reviewed - Python
Published by github-actions[bot] about 6 years ago

Larq -

:bug: Bug Fixes

Fix iteration bug in Bop with TensorFlow 2.0.0 (#286, #287) @koenhelwegen

:book: Documentation

Add explanation of scoped metric definition (#281) @leonoverweel
Reorder quantizers so they show up grouped in the docs (#274) @lgeiger

:constructionworkerman: Internal Improvements

Quantizer tests improvements (#284) @jamescook106
Run unit tests with GitHub actions (#277) @lgeiger
Only run release notes action on pushed to master (#276) @lgeiger
Add Release Drafter as GH action (#275) @lgeiger
Automatically draft release notes (#273) @lgeiger
Add github action to publish package (#272) @lgeiger
Enforce PR labels (#271) @lgeiger
Remove unnecessary github workflow file (#270) @lgeiger

:arrow_up: Dependencies

Bump vega-embed@6 (#285) @lgeiger
Bump absl-py from 0.8.0 to 0.8.1 (#283) @dependabot-preview
Bump pytest from 5.2.0 to 5.2.1 (#280) @dependabot-preview
Bump pytest-cov from 2.8.0 to 2.8.1 (#279) @dependabot-preview
Bump pytest-cov from 2.7.1 to 2.8.0 (#278) @dependabot-preview
Bump mkdocs-material from 4.4.2 to 4.4.3 (#269) @dependabot-preview

Scientific Software - Peer-reviewed - Python
Published by lgeiger about 6 years ago

Larq -

🚀 Performance Improvements

Explicitly clip gradients of quantizers in backwards pass (#267)

📖 Documentation

Extend site description (#260)
Extend example on front page and remove it from user guide (#261)
Cite BNN NeurIPS paper (#258)
Mathematically clarify binarization in guide (#259)
Add formula for swish sign and improve docs for magnitude aware sign (#257)
Add DoReFa-Net to documentation (#253)
General clarification of Key concepts page (#254)

👷 Internal Improvements

Bump pillow from 6.1.0 to 6.2.0 (#266)
Bump tensorflow used by Binder from 1.14.0 to 2.0.0 (#265)
Test agains TensorFlow 2.0.0 (#264)
Bump larq-zoo from 0.4.1 to 0.4.2 (#263)
Bump pytest from 5.1.3 to 5.2.0 (#255)

Scientific Software - Peer-reviewed - Python
Published by lgeiger over 6 years ago

Larq -

:tada: Features

Add SteHeaviside quantizer (#222)
Add Dorefa quantizer (#223)
Add option to use unclipped identity STE in quantizers (#232)

📖 Documentation

Add JOSS publication (#217)
Minor updates and clarifications (#213, #227, #233, #236, #237, #238, #239, #240, #241, #242, #244)
Optimize images (#220)
Styling updates (#219)
Use Altair/Vega to plot activations and quantizers in API docs (#231)

👷 Internal Improvements

Add LarqMetric class, add register_alias decorator (#225)
Update dependencies: (pytest 5.1.3 #248, larq-zoo 0.4.1 #214, mathjax 2 #228, mknotebooks 0.1.7 #250)
Improve tests (#216, #221, #234, #252)

Scientific Software - Peer-reviewed - Python
Published by AdamHillier over 6 years ago

Larq -

🐛 Bugs

Fix MAC count strings. (#208, #209)
Fix Tensor comparison to support TensorFlow 2.0.0-rc0 hashing behaviour (#210)

📖 Documentation

Cleanup Home page and add logo (#196)
Move larq_zoo docs to main repo (#199)
Cleanup front page (#201)
Update larq_zoo docs to the latest version (#202)
Add docs formatter to automatically add netron links (#200)
Switch hosting from GitHub Pages to Zeit Now (#204)
Unify colour scheme and fix some minor styling issues (#205)
Fix class of tab bar to correctly reflow on mobile devices (#207)

👷 Internal Improvements

Bump pytest from 5.1.1 to 5.1.2 (#197)
Bump pymdown-extensions from 6.0 to 6.1 (#198)

Scientific Software - Peer-reviewed - Python
Published by lgeiger over 6 years ago

Larq -

:tada: Features

Allow to configure gradient clipping in quantizers (#182)
- This adds class based variations of all quantizers to make gradient clipping configurable
Add SwishSign quantizer (#172)
Format numbers in larq model summaries to be more readable (#192, #193)

:bug: Bugs

Correctly display non trainable parameters of layers in summary (#180)

📖 Documentation

Use new logo and colour scheme in website (#194)

:constructionworkerman: Internal Improvements

Cleanup ternary quantizer and add alias to ste_tern` (#171)
Cleanup quantizer classes (#181)
Bump pytest from 5.0.1 to 5.1.1 (#185, #187)
Bump mkdocs-material from 4.4.0 to 4.4.2 (#188, #189)
Bump absl-py from 0.7.1 to 0.8.0 (#190)

Scientific Software - Peer-reviewed - Python
Published by lgeiger over 6 years ago

Larq -

:warning: Breaking Changes :warning:

Change default ternarization threshold in SteTern (#151)
Rename layer metrics for better grouping in TensorBoard (#165)

:tada:Features

Add MAC counter to larq.models.summmary (#157)
Add metrics scope contextmanager to easily configure training metrics (#163, #169)

:bug: Bugs

Fix RuntimeError: generator didn't yield in distributed setting (#167)

:book: Documentation

Correct inequality sign in SteTern documentation (#144)
Bump snapshottest from 0.5.0 to 0.5.1 (#145)
Improved hyperlinks, plus docs on quantizers as activations (#146)
FAQ question on open-source (#149)
Consistently name Binarized Neural Networks (#150)
Add interactive plot for building BNN guide (#152)
Add issue templates (#153)
Include xnornet plots in documentation (#154)
Chang site description (#161)
Add sticky tab navigation (#162, #164)

:constructionworkerman: Internal Improvements

Simplify layer base classes (#156)
Simplify CI config using pytest-azurepipelines (#158)
Add pyflakes and black linting (#159)

Scientific Software - Peer-reviewed - Python
Published by lgeiger over 6 years ago

Larq -

🎉 Features

Publish Larq Zoo and add pretrained models (#143) https://larq.dev/models, https://github.com/larq/zoo

📖 Documentation

Add guides, about and FAQ (#138)
Regenerate model summary in example notebooks (#134)

🐛 Bugs

Ignore metric weights in layers (#136)

👷Internal Improvements

Move to larq its own GitHub organisation (#140, #141, #142)
Only export larq.models.summary in larq.models (#135)

Scientific Software - Peer-reviewed - Python
Published by lgeiger over 6 years ago

Larq -

⚠️ Breaking Changes ⚠️

Deprecate QuantizationLogger and layers.quantized_weights (#122) This has been replaced by layer metrics introduced in #100

🎉 Features

Improve model summary and add basic profiling (#115, #121, #124, #125, #128, #132, #133)
Add layer metrics to replace quant logger callback (#100)
Add DepthwiseConv2D layer (#119)

📖 Documentation

Fix some minor typos in docs (#114)
Fix twitter link and minify generated html pages to save a few kB (#116)
Fix Binder links for example notebooks (#129)
Fix typo in SteTern documentation (#130)
Add links to notebooks on GitHub and improve buttons (#131)

👷 Internal Improvements

Add simple unit test for TF 1.14 learning rate schedules (#106, #127)

Scientific Software - Peer-reviewed - Python
Published by lgeiger over 6 years ago

Larq -

🎉 Features

Add Bop optimizer (#103)
Add larq.math.sign (#110)
Add a generic hyperparameter scheduler (#104)
Add epoch level logging to QuantizationLogger (#97)

🐛 Bugs

Prevent QuantizationLogger from adding nodes to the graph during training (#99)

📖 Documentation

Setup https://larq.dev (#113)
Add Papers using Larq page (#112)
Add docs to cleanup old build files (#109)
Change Top level docs structure (#98)

👷 Internal Improvements

Test against Python 3.6/3.7 and TF 1.13.1/1.14.0/2.0-beta1 (#107

Scientific Software - Peer-reviewed - Python
Published by lgeiger over 6 years ago

Larq -

:tada: Features

Add leaky tanh activation (#78)
Add magnitude aware sign quantizer (#79)
Add ternary quantizer (#86)

:bug: Bugs

Lazily compute quantized_weights (#96)

:construction_worker: Internal Improvements

Switch to plain mkdocs to build our documentation (#75)
Tests: Compute gradient in eager mode (#74)
Add test to make sure quantizers can be used as activations (#82)
Test QuantizationLogger with TF 2 nightly (#85)

:memo: Documentation

Add matplotlib to requirements.txt so our example can be run on binder ( … #67)
Add docs on how to publish a pip release (#68)
Advanced binarynet tutorial (#71, #72)
Rework Readme and wording in the examples (#73)
Add User Guide to docs (#80)
Plot output and gradients of activations and quantizers in docs (#81)
Improve docs on input quantizer (#84)
Display links to source code (#93)

Scientific Software - Peer-reviewed - Python
Published by lgeiger over 6 years ago

Larq -

:bug: Bugs

Require Python 3.6+ for larq to be installed #66

Scientific Software - Peer-reviewed - Python
Published by lgeiger over 6 years ago

Larq - v0.1.0

🎉 Initial Beta Release 🎉

Scientific Software - Peer-reviewed - Python
Published by lgeiger over 6 years ago

Recent Releases of Larq

Larq - v0.13.3

:bug: Bug Fixes

:arrow_up: Dependencies

Larq - v0.13.2

:tada: Features

:arrow_up: Dependencies

Larq - v0.13.1

:tada: Features

:constructionworkerman: Internal Improvements

:arrow_up: Dependencies

Larq - v0.13.0

:warning: Breaking Changes :warning:

:tada: Features

:constructionworkerman: Internal Improvements

:arrow_up: Dependencies

Larq - v0.12.2

:bug: Bug Fixes

:constructionworkerman: Internal Improvements

:arrow_up: Dependencies

Larq - v0.12.1

:bug: Bug Fixes

:book: Documentation

:constructionworkerman: Internal Improvements

:arrow_up: Dependencies

Larq - v0.12.0

:tada: Features

:book: Documentation

:constructionworkerman: Internal Improvements

:arrow_up: Dependencies

Larq - v0.11.2

:bug: Bug Fixes

Larq - v0.11.1

:bug: Bug Fixes

:constructionworkerman: Internal Improvements

Larq - v0.11.0

:warning: Breaking Changes :warning:

:bug: Bug Fixes

:constructionworkerman: Internal Improvements

:arrow_up: Dependencies

Larq - v0.10.3

:bug: Bug Fixes

:book: Documentation

:constructionworkerman: Internal Improvements

:arrow_up: Dependencies

Larq - v0.10.2

:constructionworkerman: Internal Improvements

:arrow_up: Dependencies

Larq -

:bug: Bug Fixes

:book: Documentation

:constructionworkerman: Internal Improvements

:arrow_up: Dependencies

Larq - v0.10.0

:tada: Features

:bug: Bug Fixes

:book: Documentation

:constructionworkerman: Internal Improvements

:arrow_up: Dependencies

Larq - v0.9.6

:bug: Bug Fixes

:book: Documentation

:arrow_up: Dependencies

Larq -

:bug: Bug Fixes

:book: Documentation

:arrow_up: Dependencies

Larq -

:tada: Features

:book: Documentation

:constructionworkerman: Internal Improvements

:arrow_up: Dependencies

Larq -

:bug: Bug Fixes

Larq - v0.9.2

:bug: Bug Fixes

:arrow_up: Dependencies

Larq -

:bug: Bug Fixes

Larq -