Skip to content

Latest commit

 

History

History
65 lines (48 loc) · 1.91 KB

CHANGELOG.md

File metadata and controls

65 lines (48 loc) · 1.91 KB

Change Log

Version 0.2.1

Areas of improvement
  • Added set_regularization function
  • Added beta argument to dice loss
  • Added threshold argument for metrics
  • Fixed prerprocess_input for mobilenets
  • Fixed missing parameter interpolation in ResizeImage layer config
  • Some minor improvements in docs, fixed typos

Version 0.2.0

Areas of improvement
  • New backbones (SE-ResNets, SE-ResNeXts, SENet154, MobileNets)
  • Metrcis:
    • iou_score / jaccard_score
    • f_score / dice_score
  • Losses:
    • jaccard_loss
    • bce_jaccard_loss
    • cce_jaccard_loss
    • dice_loss
    • bce_dice_loss
    • cce_dice_loss
  • Documentation Read the Docs
  • Tests + Travis-CI
API changes
  • Some parameters renamed (see API docs)
  • encoder_freeze=True does not freeze BatchNormalization layer of encoder
Thanks

@IlyaOvodov #15 #37 investigation of align_corners parameter in ResizeImage layer
@NiklasDL #29 investigation about convolution kernel in PSPNet final layers

Version 0.1.2

Areas of improvement
  • Added PSPModel
  • Prepocessing functions for all backbones:
from segmentation_models.backbones import get_preprocessing

preprocessing_fn = get_preprocessing('resnet34')
X = preprocessing_fn(x)
API changes
  • Default param use_batchnorm=True for all decoders
  • FPN model Upsample2D layer renamed to ResizeImage

Version 0.1.1

  • Added Linknet model
  • Keras 2.2+ compatibility (fixed import of _obtain_input_shape)
  • Small code improvements and bug fixes

Version 0.1.0

  • Unet and FPN models