Skip to content

Latest commit

 

History

History
383 lines (215 loc) · 9.37 KB

CHANGELOG.md

File metadata and controls

383 lines (215 loc) · 9.37 KB
Raw

0.1.25 — 2024-07-05

Features

  • Image encoders are imported now only from timm models.
  • Add enc_out_indices to model classes, to enable selecting which layers to use as the encoder outputs.

Removed

  • Removed SAM and DINOv2 original implementation image-encoders from this repo. These can be found from timm models these days.
  • Removed cellseg_models_pytorch.training module which was left unused after example notebooks were updated.

Examples

  • Updated example notebooks.
  • Added new example notebooks utilizing UNI foundation model from the MahmoodLab.
  • Added new example notebooks utilizing the Prov-GigaPath foundation model from the Microsoft Research.
  • NOTE: These examples use the huggingface model hub to load the weights. Permission to use the model weights is required to run these examples.

Chore

  • Update timm version to above 1.0.0.

Breaking changes

  • Lose support for python 3.9
  • The self.encoder in each model is new, thus, models with trained weights from previous versions of the package will not work with this version.

0.1.24 — 2023-10-13

Style

  • Update the Ìnferer.infer() -method api to accept arguments related to saving the model outputs.

Features

  • Add CPP-Net. https://arxiv.org/abs/2102.06867

  • Add option for mixed precision inference

  • Add option to interpolate model outputs to a given size to all of the segmentation models.

  • Add DINOv2 Backbone

  • Add support for .geojson, .feather, .parquet file formats when running inference.

Docs

  • Add CPP-Net example trainng with Pannuke dataset.

Fixes

  • Fix resize transformation bug.

0.1.23 — 2023-08-28

Features

  • add a stem-skip module. (Long skip for the input image resolution feature map)

  • add UnetTR transformer encoder wrapper class

  • add a new Encoder wrapper for timm and unetTR based encoders

  • Add stem skip support and upsampling block options to all current model architectures

  • Add masking option to all the criterions

  • Add MAELoss

  • Add BCELoss

  • Add base class for transformer based backbones

  • Add SAM-VitDet image encoder with support to load pre-trained SAM weights

  • Add CellVIT-SAM model.

Docs

  • Add notebook example on training Hover-Net with lightning from scratch.

  • Add notebook example on training StarDist with lightning from scratch.

  • Add notebook example on training CellPose with accelerate from scratch.

  • Add notebook example on training OmniPose with accelerate from scratch.

  • Add notebook example on finetuning CellVIT-SAM with accelerate.

Fixes

  • Fix current TimmEncoder to store feature info

  • Fix Up block to support transconv and bilinear upsampling and fix data flow issues.

  • Fix StardistUnet class to output all the decoder features.

  • Fix Decoder, DecoderStage and long-skip modules to work with up scale factors instead of output dimensions.

0.1.22 — 2023-07-10

Features

  • Add mps (Mac) support for inference
  • Add cell class probabilities to saved geojson files

0.1.21 — 2023-06-12

Features

Fixes

  • Minor bug fixes

0.1.20 — 2023-01-13

Fixes

  • Enable writing folder & hdf5 datasets with only images (previously needed image-mask pairs)

  • Enable writing datasets without patching.

  • Add long missing h5 reading utility function to FileHandler

Features

  • Add hdf5 input file reading to Inferer classes.

  • Add option to write pannuke dataset to h5 db in PannukeDataModule and LizardDataModule.

  • Add a generic model builder function get_model to models.__init__.py

  • Rewrite segmentation benchmarker. Now it can take in hdf5 datasets.

0.1.19 — 2023-01-04

Features

  • Add pytorch lightning in-built auto_lr_finder option to SegmentationExperiment

0.1.18 — 2023-01-03

Features

  • Add Multi-scale-convolutional-attention (MSCA) module (SegNexT).
  • Add TokenMixer & MetaFormer modules.

0.1.17 — 2022-12-29

Features

  • Add transformer modules
  • Add exact, slice, and memory efficient (xformers) self attention computations
  • Add transformers modules to Decoder modules
  • Add common transformer mlp activation functions: star-relu, geglu, approximate-gelu.
  • Add Linformer self-attention mechanism.
  • Add support for model intialization from yaml-file in MultiTaskUnet.
  • Add a new cross-attention long-skip module. Works with long_skip='cross-attn'

Refactor

  • Added more verbose error messages for the abstract wrapper-modules in modules.base_modules
  • Added more verbose error catching for xformers.ops.memory_efficient_attention.

0.1.16 — 2022-12-14

Dependencies

  • Bump old versions of numpy & scipy

0.1.15 — 2022-12-13

Features

  • Use the inferer class as input to segmentation benchmarker class

0.1.14 — 2022-12-01

Performance

  • Throw away some unnecessary parts of the cellpose post-proc pipeline that just brought overhead and did nothing.

Refactor

  • Refactor the whole cellpose post-processing pipeline for readability.

  • Refactored multiprocessing code to be reusable and moved it under utils.

Features

  • Add exact euler integration (on CPU) for cellpose post-processing.

  • added more pathos.Pool options for parallel processing. Added ThreadPool, ProcessPool & SerialPool

  • add all the mapping methods for each Pool obj. I.e. amap, imap, uimap and map

Tests

  • added tests for the multiprocessing tools.

0.1.13 — 2022-11-25

Features

  • Add option to return encoder features, and decoder features along the outputs in the forward pass of any model.

Fixes

  • Turn the cellpose and stardist postproc dirs into modules.

0.1.12 — 2022-11-03

Performance

  • Reverse engineered the stardist post-processing pipeline to python. Accelerated it with Numba and optimized it even further. Now it runs almost 2x faster than the original C++ verion.

Fixes

  • Fix bug with padding in SlidingWindowInferer

0.1.11 — 2022-10-21

Removed

  • unnecessary torchvision dependency

0.1.10 — 2022-10-21

Removed

  • torch-optimizer removed from the optional dependency list. Started to cause headache.

0.1.9 — 2022-10-21

Refactor

  • Moved saving utilities to FileHandler and updated tests.

Features

  • Added geojson saving support for inference

0.1.8 — 2022-10-18

Features

  • Support to return all of the feature maps from each decoder stage.

  • Add multi-gpu inference via DataParallel

0.1.7 — 2022-10-15

Fixes

  • Fix SCE loss bug.

0.1.6 — 2022-10-14

Features

  • Add a Wandb artifact table callback for loading a table of test data metrics and insights to wandb.

Fixes

  • Symmetric CE loss fixed.

  • Add option to return binary and instance labelled mask from the dataloader. Previously binary was returned with return_inst flag which was confusing.

  • Fix the SegmentationExperiment to return preds and masks at test time.

0.1.5 — 2022-10-07

Fixes

  • Wandb Callback bugs fixed.

0.1.4 — 2022-10-06

Test

  • Update loss tests

Fixes

  • Add a conv block BasicConvOld to enable Dippa to cellseg conversion of models.

  • Fix inst_key, aux_key bug in MultiTaskUnet

  • Add a type_map > 0 masking for the inst_maps in post-processing

  • Modify the optimizer adjustment utility function to adjust any optim/weight params.

  • Modify lit SegmentationExperiment according to new changes.

Features

  • Add optional spectral decoupliing to all losses

  • Add optional Label smoothing to all losses

  • Add optional Spatially varying label smoothing to all losses

  • Add mse, ssim and iqi torchmetrics for metric logging.

  • Add wandb per class metric callback for logging.

  • Add from_yaml init classmethod to initialize from yaml files.

0.1.3 — 2022-09-23

Test

  • Update tests for Inferes and mask utils.
  • Add tests for the benchmarkers.

Fixes

  • init and typing fixes

Docs

  • Typo fies in docs

Features

  • Add numba parallellized median filter and majority voting for post-processing

  • Add support for own semantic and type seg post-proc funcs in Inferers

  • Add segmentation performance benchmarking helper class.

  • Add segmentation latency benchmarking helper class.

0.1.2 — 2022-09-09

Fixes

  • Update save2db & save2folder for optional type_map and sem_map args.
  • Pre-processing (pre-proc) callable arg for _get_tiles method. This enables the Lizard datamodule.
  • Fix- padding bug with sliding window inference.

Features

Test

  • Update dataset tests.

  • Update tests for multi-task U-Net

Type Hints

  • Fix incorrect type hints.

Examples

  • Add cellpose training with Lizard dataset notebook.