georgemilosh/Climate-Learning

Issues

• Restrict labels to heat wave onsets

  #18 opened 3 years ago by georgemilosh
  2

• Multiclass classification

  #12 opened 3 years ago by AlessandroLovo
  0

• Try subtracting the average yearly temperature for computing the labels

  #25 opened 3 years ago by georgemilosh
  2

• Fix the bug with scheduler_kwargs not showing up in config file

  #26 opened 2 years ago by georgemilosh
  2

• xarray adaptations in Learn2_New

  #29 opened 2 years ago by AlessandroLovo
  1

• xarray masks

  #28 opened 2 years ago by AlessandroLovo
  1

• xarray Plasim Field

  #27 opened 2 years ago by AlessandroLovo
  1

• mrso masking bug

  #19 opened 2 years ago by AlessandroLovo
  4

• early stopping for eons

  #17 opened 3 years ago by AlessandroLovo
  0

• Xarrays

  #9 opened 3 years ago by AlessandroLovo
  2

• Support for padding that respects the topology of the sphere

  #24 opened 3 years ago by georgemilosh
  0

• Add possibility of continuing runs

  #21 opened 3 years ago by AlessandroLovo
  0

• Add possibility to train a model without t2m field that is used as a label

  #22 opened 3 years ago by georgemilosh
  0

• Change the skip connection structure from dictionary to a list

  #23 opened 3 years ago by georgemilosh
  0

• Accelerate the VAE routines

  #20 opened 3 years ago by georgemilosh
  0

• Transfer learning for resampling

  #16 opened 3 years ago by AlessandroLovo
  1

• Use Learn2_new.py for vae_learn2.py

  #13 opened 3 years ago by georgemilosh
  0

• More modular transfer learning

  #15 opened 3 years ago by AlessandroLovo
  1

• Fix errors with resampling.py

  #14 opened 3 years ago by georgemilosh
  1

• Proper plotting routines for rolled arrays

  #10 opened 3 years ago by georgemilosh
  2

• Better transfer learning

  #8 opened 3 years ago by AlessandroLovo
  1

• Sign of tau

  #11 opened 3 years ago by AlessandroLovo
  2

• Early stopping

  #7 opened 3 years ago by AlessandroLovo
  1

• :sparkles: **Modular support for PLASIM/Learn2.py** :sparkles:

  #1 opened 3 years ago by georgemilosh
  2

• Discrete metric computed on the fly don't unbias the probabilities

  #6 opened 3 years ago by AlessandroLovo
  2

• Better logging

  #2 opened 3 years ago by AlessandroLovo
  1

• Efficient computations over multiple runs

  #4 opened 3 years ago by AlessandroLovo
  1

• utilities

  #5 opened 3 years ago by AlessandroLovo
  1

• runs.json

  #3 opened 3 years ago by AlessandroLovo
  1