Skip to content

Latest commit

 

History

History
33 lines (27 loc) · 1.33 KB

updates.md

File metadata and controls

33 lines (27 loc) · 1.33 KB
Raw

Update

Version 2.2.0

  • MOFTransformer can now proceed with multi-task learning.
  • Add n_targets in config (multi-task learning)

Version 2.1.5

  • transformer < 4.39.0 in requirements

Version 2.1.4

  • Modify README.md
  • Modify the setup of datamodules in the predict function to only proceed with the corresponding split

Version 2.1.3

  • Add "max_supercell_atoms" in config

Version 2.1.2

  • Check GRIDAY is vailed or not when you run install_griday
  • torchmetrics < 1.0.0 in requirements
  • Fix bugs in predict when loss is classification and n_classes are 2.

Version 2.1.1

  • Fixed a bug when the structure name of raw_[downstream].json contains a cif during prepare_data.
  • Changed an error that occurred when there were multiple devices in an interactive environment to a warning, automatically converting the configuration to a single device.
  • Changed a bug for command-line interface (moftransformer run)
  • Change float type 16 -> 32 before denormalization.
  • Add command-line interface (moftransformer predict)
  • Modified minor typo errors in document
  • Remove logger file in function "predict"
  • Make "test" function
  • Fix 'predict' function in classification task -> add "classification_logits_index'
  • 'best.ckpt' are generated after training
  • Update Readme.md to include the comment regarding only support for Linux