sinzlab/nnfabrik

Issues

• Argument to turn off strict model state dict loading checks

  #148 opened 10 months ago by MaxFBurg
  0

• not compatible version for "ax", "botorch" and "gpytorch" if just install ax-platform

  #143 opened 2 years ago by ShanqMa
  0

• Not compatible with Python3.10

  #142 opened 2 years ago by pollytur
  1

• NNFabrik uses PyTorch but doesn't specify it as a requirement

  #129 opened 4 years ago by christoph-blessing
  0

• recommend mysql settings

  #128 opened 4 years ago by Matthijspals
  0

• Missing exception in try and except block

  #126 opened 4 years ago by christoph-blessing
  0

• nnfabrik.schema_name deprecated but required without access to nnfabrik_core

  #125 opened 4 years ago by chbehrens
  2

• KeyError in TrainedModel when no data_info_table is defined

  #124 opened 4 years ago by chbehrens
  0

• Fix the bug in DataInfoBase table

  #115 opened 4 years ago by mohammadbashiri
  1

• Readme not up-to-date

  #118 opened 4 years ago by ArneNx
  0

• Remove reference to neuralpredictors

  #112 opened 4 years ago by ArneNx
  2

• Hypersearch does not work with `my_nnfabrik`

  #111 opened 4 years ago by KonstantinWilleke
  0

• loading a model fails when `include_dataloaders=False` because there is no DataInfo table

  #102 opened 4 years ago by mohammadbashiri
  5

• Centralized caching of checkpoints during training

  #56 opened 4 years ago by ArneNx
  1

• Minimal example of SinzLab use-case

  #81 opened 4 years ago by mohammadbashiri
  6

• Add schema creation to the example notebook

  #87 opened 4 years ago by mohammadbashiri
  1

• Documentation of tables and functions + Best Practice

  #25 opened 5 years ago by KonstantinWilleke
  2

• Decouple measure tables (per data_key local identity) and neuron id (global identity)

  #75 opened 4 years ago by mohammadbashiri
  2

• Move out Code for Dynamic Models

  #59 opened 4 years ago by fabiansinz
  0

• Move out Monkey Code

  #58 opened 4 years ago by fabiansinz
  0

• Cleaning Up before going public

  #40 opened 4 years ago by KonstantinWilleke
  3

• Endpoints and keys in datajoint config are allowed to be undefined

  #93 opened 4 years ago by linuswagner
  0

• Keyword argument processing and parameter optimization in early_stopping_trainer

  #20 opened 4 years ago by KonstantinWilleke
  0

• Update the Trainer to make use of Pytorch Lightning

  #22 opened 4 years ago by mohammadbashiri
  3

• Throw error if current user is not found

  #52 opened 5 years ago by christoph-blessing
  0

• Error when there are more than two variables return by dataloader

  #62 opened 4 years ago by mohammadbashiri
  1

• Discourage saving the model's state dict for pretrained models.

  #70 opened 4 years ago by KonstantinWilleke
  1

• Make Hypersearch module deal with more than one seed

  #85 opened 4 years ago by KonstantinWilleke
  0

• Toy Trainer Expects Separate Dataloaders But Gets Dictionary

  #66 opened 4 years ago by christoph-blessing
  1

• Extend README/documentation by datatypes

  #82 opened 4 years ago by linuswagner
  2

• Update example notebook to use templates and newest version

  #46 opened 5 years ago by ArneNx
  2

• Refactor

  #24 opened 5 years ago by mohammadbashiri
  0

• Write tests (instantiating tables and populating TrainedModel)

  #47 opened 5 years ago by ArneNx
  0

• Syntax Error with Python 3.7

  #35 opened 5 years ago by ArneNx
  1

• Introduce input arguments to make choice of losses (e.g., averaged vs not) configurable by the user

  #30 opened 5 years ago by mohammadbashiri
  1

• Make `standard_early_stop_trainer` to accept similar input arguments as the existing trainer.

  #29 opened 5 years ago by mohammadbashiri
  1

• Consistency in naming convention

  #17 opened 5 years ago by mohammadbashiri
  4

• Proposal: Abstract classes for datasets, models and trainers

  #23 opened 5 years ago by ArneNx
  3

• Separate out the trained model as a part table

  #3 opened 5 years ago by eywalker
  1