neural-fortran - neural-fortran-0.22.0

What's Changed

• Generic conv & maxpool by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/220
• Apply optimizer to model weights without data copy by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/222
• Rename locallyconnected1d -> locallyconnected2d by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/224

Full Changelog: https://github.com/modern-fortran/neural-fortran/compare/v0.21.0...v0.22.0

- Fortran
Published by milancurcic 11 months ago

neural-fortran - neural-fortran-0.21.0

What's Changed

• Generic reshape by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/213
• Multihead Attention Fixes by @OneAdder in https://github.com/modern-fortran/neural-fortran/pull/209

Full Changelog: https://github.com/modern-fortran/neural-fortran/compare/v0.20.0...v0.21.0

- Fortran
Published by milancurcic about 1 year ago

neural-fortran - neural-fortran-0.20.0

What's Changed

• Layernorm by @OneAdder in https://github.com/modern-fortran/neural-fortran/pull/203
• Embedding Layer by @OneAdder in https://github.com/modern-fortran/neural-fortran/pull/205
• Locally connected layer by @ricor07 in https://github.com/modern-fortran/neural-fortran/pull/201

New Contributors

• @ricor07 made their first contribution in https://github.com/modern-fortran/neural-fortran/pull/201

Full Changelog: https://github.com/modern-fortran/neural-fortran/compare/v0.19.0...v0.20.0

- Fortran
Published by milancurcic over 1 year ago

neural-fortran - neural-fortran-0.19.0

What's Changed

• Update for cmake use of neural-fortran by @mathomp4 in https://github.com/modern-fortran/neural-fortran/pull/192
• Add Input2d layer by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/198
• Generic flatten (2d and 3d) by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/202
• Linear2d layer by @OneAdder in https://github.com/modern-fortran/neural-fortran/pull/197
• Dropout layer by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/194
• Multihead attention by @OneAdder in https://github.com/modern-fortran/neural-fortran/pull/199

New Contributors

• @mathomp4 made their first contribution in https://github.com/modern-fortran/neural-fortran/pull/192
• @OneAdder made their first contribution in https://github.com/modern-fortran/neural-fortran/pull/197

Full Changelog: https://github.com/modern-fortran/neural-fortran/compare/v0.18.0...v0.19.0

- Fortran
Published by milancurcic over 1 year ago

neural-fortran - neural-fortran-0.18.0

What's Changed

• Remove Keras HDF5 by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/188
• Standard conformance fixes by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/190

The Keras HDF5 loading feature is now in neural-fortran/nf-keras-hdf5.

Full Changelog: https://github.com/modern-fortran/neural-fortran/compare/v0.17.0...v0.18.0

- Fortran
Published by milancurcic almost 2 years ago

neural-fortran - neural-fortran-0.17.0

What's Changed

• Proposition of API for the method network % evaluate by @jvdp1 in https://github.com/modern-fortran/neural-fortran/pull/182
• Remove some unused variables and a broadcast by @jvdp1 in https://github.com/modern-fortran/neural-fortran/pull/180
• Intrinsic pack replaced by pointers in get_params and get_gradients by @jvdp1 in https://github.com/modern-fortran/neural-fortran/pull/183

Full Changelog: https://github.com/modern-fortran/neural-fortran/compare/v0.16.1...v0.17.0

- Fortran
Published by milancurcic about 2 years ago

neural-fortran - neural-fortran-0.16.1

What's Changed

• Addidtion of Findneural-fortran.cmake by @jvdp1 in https://github.com/modern-fortran/neural-fortran/pull/178
• Replacement of a matmul + use of merge by @jvdp1 in https://github.com/modern-fortran/neural-fortran/pull/181

Full Changelog: https://github.com/modern-fortran/neural-fortran/compare/v0.16.0...v0.16.1

- Fortran
Published by milancurcic about 2 years ago

neural-fortran - neural-fortran-0.16.0

What's Changed

• More conv2d tests by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/174
• Add instructions using Conda by @certik in https://github.com/modern-fortran/neural-fortran/pull/172
• Addition of the Loss derived type and of the MSE loss function by @jvdp1 in https://github.com/modern-fortran/neural-fortran/pull/175

New Contributors

• @certik made their first contribution in https://github.com/modern-fortran/neural-fortran/pull/172

Full Changelog: https://github.com/modern-fortran/neural-fortran/compare/v0.15.1...v0.16.0

- Fortran
Published by milancurcic about 2 years ago

neural-fortran - neural-fortran-0.15.1

What's Changed

• safegaurd Box-Muller normal random number generation against u=0.0 by @dacarnazzola in https://github.com/modern-fortran/neural-fortran/pull/158

New Contributors

• @dacarnazzola made their first contribution in https://github.com/modern-fortran/neural-fortran/pull/158

Full Changelog: https://github.com/modern-fortran/neural-fortran/compare/v0.15.0...v0.15.1

- Fortran
Published by milancurcic almost 3 years ago

neural-fortran - neural-fortran-0.15.0

What's Changed

• Adagrad Optimizer Implementation by @Spnetic-5 in https://github.com/modern-fortran/neural-fortran/pull/154

Full Changelog: https://github.com/modern-fortran/neural-fortran/compare/v0.14.0...v0.15.0

- Fortran
Published by Spnetic-5 almost 3 years ago

neural-fortran - neural-fortran-0.14.0

What's Changed

• Added Adam optimizer implementation by @Spnetic-5 in https://github.com/modern-fortran/neural-fortran/pull/150

Full Changelog: https://github.com/modern-fortran/neural-fortran/compare/v0.13.0...v0.14.0

- Fortran
Published by Spnetic-5 almost 3 years ago

neural-fortran - neural-fortran-0.13.0

What's Changed

• Contributing guide by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/109
• GSoC Optimizers: Example program to fit a quadratic function by @Spnetic-5 in https://github.com/modern-fortran/neural-fortran/pull/134
• Update contributors and acknowledgment by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/138
• Add missing update for conv2d_layer by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/141
• Example program that shows how to access internal layer parameters by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/140
• add CELU activation function by @pablomazo in https://github.com/modern-fortran/neural-fortran/pull/143
• Added RMSProp Optimizer subroutine by @Spnetic-5 in https://github.com/modern-fortran/neural-fortran/pull/144
• SGD optimizer stub by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/139
• Connect flatten, conv2d, and maxpool2d layers in backward pass by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/142
• Added Momentum and Nesterov modifications by @Spnetic-5 in https://github.com/modern-fortran/neural-fortran/pull/148

New Contributors

• @pablomazo made their first contribution in https://github.com/modern-fortran/neural-fortran/pull/143

Full Changelog: https://github.com/modern-fortran/neural-fortran/compare/v0.12.0...v0.13.0

- Fortran
Published by Spnetic-5 almost 3 years ago

neural-fortran - neural-fortran-0.12.0

What's Changed

• Added Leaky ReLU activation function (1D & 3D) by @Spnetic-5 in https://github.com/modern-fortran/neural-fortran/pull/123
• Class-based activation functions by @ggoyman in https://github.com/modern-fortran/neural-fortran/pull/126

New Contributors

• @Spnetic-5 made their first contribution in https://github.com/modern-fortran/neural-fortran/pull/123
• @ggoyman made their first contribution in https://github.com/modern-fortran/neural-fortran/pull/126

Full Changelog: https://github.com/modern-fortran/neural-fortran/compare/v0.11.0...v0.12.0

- Fortran
Published by milancurcic about 3 years ago

neural-fortran - neural-fortran-0.11.0

What's Changed

• Add missing edit descriptor by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/113
• Small workaround to make Intel compiler happy by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/117
• Introduce optimizerbasetype in support of different optimizers by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/116
• Insert a flatten layer if a dense layer follows a layer by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/118
• Add linear activation function by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/119
• Update contributors by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/120

Full Changelog: https://github.com/modern-fortran/neural-fortran/compare/v0.10.0...v0.11.0

- Fortran
Published by milancurcic over 3 years ago

neural-fortran - neural-fortran-0.10.0

This release introduces get_num_params, get_params, and set_params methods to network and layer derived types, and allow you to more easily get and set network hyperparameters from custom Fortran code or other libraries. See the example to learn how it works.

Thanks to Christopher Zapart @jvo203 for this feature contribution.

What's Changed

• Fix CMake build by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/110
• Get set network parameters by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/111

Full Changelog: https://github.com/modern-fortran/neural-fortran/compare/v0.9.0...v0.10.0

- Fortran
Published by milancurcic over 3 years ago

neural-fortran - neural-fortran-0.9.0

This release introduces the backward passes for the conv2d and maxpool2d layers and enables the training of convolutional networks.

What's Changed

• CNN backward pass by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/99

Full Changelog: https://github.com/modern-fortran/neural-fortran/compare/v0.8.0...v0.9.0

- Fortran
Published by milancurcic over 3 years ago

neural-fortran - neural-fortran-0.8.0

This release introduces the reshape layer for connecting rank-1 layers to rank-3 layers, including the capability to read Keras's Reshape layer from h5 files.

What's Changed

• Bump version in fpm.toml by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/96
• Implement the reshape layer by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/97

Full Changelog: https://github.com/modern-fortran/neural-fortran/compare/v0.7.0...v0.8.0

- Fortran
Published by milancurcic over 3 years ago

neural-fortran - neural-fortran-0.7.0

What's Changed

• CI by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/87
• Batch inference by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/90
• Rename output -> predict for consistency with Keras by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/92
• Update README by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/93
• Fix accidental TOC reorder from a previous PR by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/94

Full Changelog: https://github.com/modern-fortran/neural-fortran/compare/v0.6.0...v0.7.0

- Fortran
Published by milancurcic over 3 years ago

neural-fortran - neural-fortran-0.6.0

What's Changed

• Update fpm instructions in light of the required HDF5 dependency by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/80
• Add CITATION.cff by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/81
• fixed a memory leak when reading the JSON by @jacobwilliams in https://github.com/modern-fortran/neural-fortran/pull/83
• Update contributors list by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/84
• Read Conv2D, MaxPooling2D, and Flatten layers from Keras by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/85

New Contributors

• @jacobwilliams made their first contribution in https://github.com/modern-fortran/neural-fortran/pull/83

Full Changelog: https://github.com/modern-fortran/neural-fortran/compare/v0.5.0...v0.6.0

- Fortran
Published by milancurcic almost 4 years ago

neural-fortran - neural-fortran-0.5.0

What's Changed

• Support for loading Keras models by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/79. Many thanks to @scivision for a complete CMake overhaul and for assisting with adding h5fortran as a dependency. This is an experimental and minimally tested feature, supporting only input and dense layers saved in a Keras HDF5 file. See the mnistfromkeras example to see how it works.
• HDF5, h5fortran, and json-fortran are now required dependencies. HDF5 you have to provide to the build system. The latter two are taken care of automatically.

Full Changelog: https://github.com/modern-fortran/neural-fortran/compare/v0.4.0...v0.5.0

- Fortran
Published by milancurcic about 4 years ago

neural-fortran - neural-fortran-0.4.0

What's Changed

• Add note about downloading MNIST data without curl by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/67
• Forward pass for the conv2d layer by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/65
• Forward pass for a max-pooling layer by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/66
• Fix layers summary table by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/70
• fix #72 - use dir tree to expose user API by @rouson in https://github.com/modern-fortran/neural-fortran/pull/74
• Clean up example and add a note to emphasize the user API by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/76
• Fix CMake build for the new src directory structure by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/77
• Implement a flatten layer by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/75

Full Changelog: https://github.com/modern-fortran/neural-fortran/compare/v0.3.0...v0.4.0

- Fortran
Published by milancurcic about 4 years ago

neural-fortran - neural-fortran-0.3.0

What's Changed

• Refactor for convnets by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/58

Full Changelog: https://github.com/modern-fortran/neural-fortran/compare/v0.2.0...v0.3.0

- Fortran
Published by milancurcic about 4 years ago

neural-fortran - neural-fortran-0.2.0

What's Changed

• Refactor: move procedure definitions submodules by @rouson in https://github.com/modern-fortran/neural-fortran/pull/51
• Delete CAF C preprocessor macro by @rouson in https://github.com/modern-fortran/neural-fortran/pull/52
• Download mnist.tar.gz if it is missing by @rouson in https://github.com/modern-fortran/neural-fortran/pull/55
• Remove instructions about downloading MNIST data by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/56
• Workaround for submodule-related bug in GFortran-9 by @milancurcic in https://github.com/modern-fortran/neural-fortran/pull/59

Full Changelog: https://github.com/modern-fortran/neural-fortran/compare/v0.1.0...v0.2.0

- Fortran
Published by milancurcic about 4 years ago

neural-fortran - neural-fortran-0.1.0

This release is a snapshot of neural-fortran before the refactor to submodules (see #51), which raises the requirement for compiler versions.

- Fortran
Published by milancurcic about 4 years ago