Version 1.0.2 Release

• Format
• Communicator utils: barrier and abort
• Change Dockerfile to Use Build arg instead of generating interactively.

Version 1.0.1 Release

• C++ CLI for CUDA extension
• FP16 usage doc
• Update Dockerfiles and doc for v1
• Write function implementation document.

Version 1.0.0 Release

• Feature/20180622 file format converter
• Workaround cublasGemmStridedBatchedEx
• Follow-up "Re-design graph engine for rewiring"

Version 1.0.0rc2 Release

• Dynamic loss scaling for the mixed precision training
• FFT and IFFT
• ext-cuda build manylinux wheel
• document fix
• Reshape with not inplace option

Version 1.0.0pre1 Release

• Feature/20180613 rename ext cuda wheel
• All-reduce during backward
• Misc build 20180618
• Feature/20180220 cpp training mnist training
• Feature/20180428 pad function
• fix document
• add ceil and floor func to cuda ext
• Fix/20180525 fix build
• Trigonometric Functions
• Big change around build system.
• Change NULL to nullptr
• Correspond to termination of python3 support of futures library.
• Communicator with Half Precision
• [generator] Function templates uses write-only flag in cast
• Function: ClipByNom
• AMSGRAD solver
• Add NOTICE file.
• ClipByValue and ClipGradByValue functions

Version 0.9.9 Release

• [BIG UPDATE] New code geneartor and type specialization (float16 support with TensorCore)
• Hotfix/20180228 out of place division

Version 0.9.8 Release

• Add ND Convolution and Deconvolution in CUDNN
• Move build doc to GitHub markdown
• Add implementation of depthwise deconvolution.
• Configurable CUDNN workspace memory size for Convolutions
• Format.
• Enable binary weight functions to run on CUDA.
• New collectives and MPI-like group concept
• Feature/20171225 round function
• Add all_reduce interface
• auto-format

Version 0.9.7

• [build] install command for libnnabla_cuda.so and headers
• Fix Python wheel target dependencies.

Version 0.9.6

• Data Parallel Distributed Training over Multiple Nodes
• Find Volta
• More efficient workspace memory allocation than cached array
• Add implementation of depthwise convolution.

Version 0.9.5

• Add the Swish activation function
• Fix driver shutting down issue on CUDA
• Add the SELU activation function
• Feature/20171011 out of place allreduce
• LeakyReLU CUDA
• Feature/20170908 quantize
• Fix INQ on Windows
• Feature/20171010 fix build and get coverage
• Add CUDA implementation of INQAffine and INQConvolution
• Remove warning from function generator
• [c++][function] Support broadcasting scalar
• Feature/20170831 separate cpp build
• remove obsolete copyright notice
• Function: matrix_diag and matrix_diag_part
• Function: Batch sample-wise matrix multiplication
• Improve backprop options of batch_normalization with CUDNN