FOXES / forecasting

Commit History

3 days ago

  1. Refactor code structure for improved readability and maintainability

    958a339
  2. Refactor code structure and remove redundant code blocks for improved readability and maintainability. Added Sphinx API documentation

    4672e76

20 days ago

  1. first commit for the FOXES package. Added ability to import FOXES as import foxes... you can download the aia stack for a particular time and use .predict to get flux map and uncertainty maps

    f716828

23 days ago

  1. updated evaluation outputs

    5e4afb8
  2. readme and config path updates

    b63594c
  3. added bias term to loss function to encourage slightly less mse but better overalll performance across classes...

    7ef80c7
  4. merging branches

    615bcaf

  1. bug fixes for callbacks / training

    7861620

  1. updated configs

    15f7f99

  1. updated config defaults

    6b21a08

  1. improve modularity of the model

    d91774f
  2. refactiring rest of code base and adding checkpoints

    0f2e7d7
  3. continuing to simplify code so that it is easier to work with. Removing extras...

    1be6930

  1. simplifying code structure

    44e0c37

  1. cleaning foxes config

    5aa1986
  2. updated config

    ee905cf
  3. - Modified inference.py to streamline dataset loading and configuration handling.⁠

    b2e46fd
  4. foxes-lite update... made for easy inference and evaluation.... removed data preprocessing and downloading pipeline pieces.

    d707d73

  1. updated FOXES model file so that the attention mask can be local, inverted local, or global attention!

    cfbde59

  1. slight bug fixes!

    cadc779

  1. refactor configuration files and update paths

    e373c2c
  2. add flux map analysis script and configuration for detecting and tracking active regions

    74c001a

  1. updated noise so that it scales globally (doesn't use individual std)

    690a9d1

  1. enhance multi-GPU support

    e0709d9
  2. enhance multi-GPU support

    f10382a
  3. refactor ablation configuration handling and improve DataLoader cleanup

    c55ace7
  4. refine noise application in SDOAIA_dataloader and simplify GPU usage message

    2647249
  5. enhance inference configuration for improved performance

    ad520c0
  6. trying to speed up inference!

    f13eb21
  7. trying to speed up inference!

    d716a89
  8. trying to speed up inference!

    0a770c7
  9. trying to speed up inference!

    8d4e63c
  10. trying to speed up inference!

    eb4a8e0

  1. Add ablation study and spatial performance scripts; update pipeline configuration and dataset handling

    992dd3f

  1. Fixed normalization of timestamp names

    387645c
  2. Fixed normalization of timestamp names

    cb3490c

  1. Add evaluation configuration and update inference pipeline

    d656d81
  2. Refactor pipeline configuration and update data processing scripts

    112be4b
  3. Refactor configuration parameters and update model architecture references

    5f12ea9

  1. Update learning rate scheduler parameters and clean up loss calculation notes

    a06c996

  1. Rename localpatch.yaml to train_config.yaml and update notes; modify learning rate scheduler parameters and rename attention mask creation method

    bc174cc
  2. Updated model to include inverted attention naming scheme

    d846c48

  1. Added checkpointing

    ebee871

  1. Moved around some files

    5468f89

  1. Update data processing workflows

    3f6ac4d

  1. Enhance plotting themes for poster analysis, refine regression plot scaling, and add data cleanup utilities⁠

    69f9796

  1. Implement dynamic theme-based plotting for regression output and tweak default parameters⁠

    27bdbce

  1. updates

    599facb

  1. Refactor code structure for improved readability and maintainability

    e76b4fe

  1. Refined flare analysis parameters and detection logic, including adjustments to validation thresholds for peak detection, enhanced configuration settings for baseline calculations, and improved visualization in the plotting functions. Updated YAML configuration for time range and detection parameters to optimize flare event identification.

    1cb8799