Skip to content

Releases: ChrisLisbon/TorchCNNBuilder

v0.1.5

24 Sep 09:25
8288022

Choose a tag to compare

Huge auto generated doc enhance and new examples

v0.1.4

15 Oct 17:13
1004b1e

Choose a tag to compare

New features:

  • adding the ability to work with latent space dimensions through the new LatentSpaceModule
  • adding new LatentSpaceModule block to the ForecasterBase with corresponding features
  • updating corresponding doc examples

Code refactor:

  • minor code refactoring
  • updating some doc examples

v0.1.3

19 Aug 18:35
b74b783

Choose a tag to compare

  • fixing calculating of channels in Builder using proportion

v0.1.2

26 Jul 09:58
37dfc6a

Choose a tag to compare

  • Fixing padding bug in convolution blocks of several convolution layers
  • Removing mandatory initialization of input_size in Builder (+ adding new validations for create sequence builder methods)
  • Improving typing of function parameters
  • Improving Builder class encapsulation
  • Restructuring code namespaces
  • Updating and improving doc examples

v0.1.1

15 Jul 10:19
b4f5bec

Choose a tag to compare

  • updating ForecasterBase interface
  • fixing bug with 3d decoder output in ForecasterBase
  • adding new validation functions
  • updating doc
  • updating linter script
  • renaming channel growth param value ratio -> exponential
  • adding new channel growth param value power (now by default in ForecasterBase)

v0.1.0

30 Jun 18:54
232b041

Choose a tag to compare

Code refactoring:

  • decomposition of functions into different files
  • restructuring of namespaces during import
  • adding a linter pipeline after each push action in pull-requests
  • adding a linter script for the developers
  • refactoring all code logic with linter
  • updating doc

New features:

  • new normalization in the convolution block as InstanceNormNd
  • new adaptive pooling in the decoder as AdaptiveMaxPoolNd
  • new channels counters like linear and constant
  • ascending param is replaced by channel_growth_rate which can be ratio (same as ascending = False), proportion (same as ascending = True), linear and constant
  • new 3d convolutions in ForecasterBase