Larq

:tada: Features

* Add support for multi threaded interpreter (512) lgeiger
* Add support for iterators in Interpreter.predict (511) lgeiger
* Add Python wrapper for LCE interpreter (507) lgeiger

:construction_worker_man: Internal Improvements

* Remove redundant code paths: reference BGemm, BGemm functor. (510) AdamHillier
* Make Interpreter::get_shapes and get_types more generic (509) lgeiger
* Add padding to temporary arrays to ensure we don't read beyond bounds. (505) AdamHillier
* Simplify MAKE_ZERO macro (503) lgeiger

:arrow_up: Dependencies

* Bump DoozyX/clang-format-lint-action from v0.9 to v0.10 (504) dependabot

:tada: Features
- Add DoReFa-net (75)

:bug: Bugs
- Correctly pass top argument to `decode_predictions` (73)

👷‍♀ Internal Improvements
- [ImgBot] Optimize images (74)

dorefanet-v0.1.0
Pretrained weights for DoReFa-Net with 1 bit weights and 2 bit activations

🐛 Bugs
- Bi-Real Net: Fix layer bug and point to updated pretrained weights (69, 72)

👷 Internal Improvements
- ⬆️ `zookeeper0.5.3` (70)
- ⬆️ `larq0.7.1` (71)

birealnet-v0.3.0
Pretrained weights for corrected Bi-Real net architecture (the previous version contained one additional layer by accident)

🎉 Features
- Made compatible with Zookeeper >= 0.5 (62)
- Add `ReseNetE18` and `BinaryDenseNet{28,37,45}` (60)
- Add `BinaryDenseNet37Dilated` (66)

:bug: Bugs
- Update weight-download url to point to correct repo. (63, 64)

📖 Documentation
- Use ImageNet dataset version 5.0.0 (53)
- Remove irrelevant docstrings (58)
- Cleanup docs build (67)


👷 Internal Improvements
- Bump scipy from 1.3.0 to 1.3.1 (56)

binary_densenet-v0.1.0
Pretrained weights for BinaryDenseNet

resnet_e-v0.1.0
Pretrained weights for ResNetE18

:bug: Bug Fixes

* Fix weight bitpacking which could lead to non-deterministic behaviour (377) lgeiger

🎉 Features
- Use improved preprocessing for pretrained models (52):
- All models are retrained from scratch and now exceed the accuracies claimed by the respective papers: https://larq.dev/models/#available-models
- All models include plots of the training process in the docs: https://larq.dev/models/api/

📖 Documentation
- Add missing `numpy` import (51)

xnornet-v0.2.0
Pretrained weights for XNOR-Net

birealnet-v0.2.0
Pretrained weights for Bi-Real Net

binary_alexnet-v0.2.0
Pretrained weights for Binary Alexnet