Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

PCA-Based Binning via Clustering #32

Open
wants to merge 19 commits into
base: master
Choose a base branch
from
Open

PCA-Based Binning via Clustering #32

wants to merge 19 commits into from

Commits on Aug 13, 2020

  1. Add a notebook outlining the pca_forest method. 

    - This method trains forests on the four smaller "islands" that appear in the data when reduced using PCA.
    @dylanagreen
    dylanagreen committed Aug 13, 2020
    Configuration menu
    Copy the full SHA
    1ac33a3 View commit details
    Browse the repository at this point in the history

  2. Move pca_forest to notebooks folder.

    @dylanagreen
    dylanagreen committed Aug 13, 2020
    Configuration menu
    Copy the full SHA
    40fbffa View commit details
    Browse the repository at this point in the history

  3. Add a notebook outlining the pca based clustering method.

    @dylanagreen
    dylanagreen committed Aug 13, 2020
    Configuration menu
    Copy the full SHA
    fdfa4c2 View commit details
    Browse the repository at this point in the history

Commits on Aug 24, 2020

  1. Update pca_cluster with lr annealing. 

    - Various other improvements as well, like restarts and fom
optimization.
    @dylanagreen
    dylanagreen committed Aug 24, 2020
    Configuration menu
    Copy the full SHA
    588f8c5 View commit details
    Browse the repository at this point in the history

Commits on Aug 26, 2020

  1. Add a learning rate range test to find optimal lr.

    @dylanagreen
    dylanagreen committed Aug 26, 2020
    Configuration menu
    Copy the full SHA
    3068bcc View commit details
    Browse the repository at this point in the history

Commits on Aug 27, 2020

  1. Implement range test into training.

    @dylanagreen
    dylanagreen committed Aug 27, 2020
    Configuration menu
    Copy the full SHA
    0696746 View commit details
    Browse the repository at this point in the history

  2. Make passing a learning rate optional. 

    - In the case where one is passed, it uses that.
- Otherwise range tests to find a "good" one.
    @dylanagreen
    dylanagreen committed Aug 27, 2020
    Configuration menu
    Copy the full SHA
    16c2edb View commit details
    Browse the repository at this point in the history

Commits on Aug 31, 2020

  1. Properly implement the range test for lr.

    @dylanagreen
    dylanagreen committed Aug 31, 2020
    Configuration menu
    Copy the full SHA
    7aa4243 View commit details
    Browse the repository at this point in the history

  2. Add notebook outlining binning as clustering method.

    @dylanagreen
    dylanagreen committed Aug 31, 2020
    Configuration menu
    Copy the full SHA
    05be1e4 View commit details
    Browse the repository at this point in the history

  3. Correctly implement range test. 

    - Notebook is a bit messy right now because I'm working fast and late at
night.
    @dylanagreen
    dylanagreen committed Aug 31, 2020
    Configuration menu
    Copy the full SHA
    f43057b View commit details
    Browse the repository at this point in the history

  4. Add script version of pca_clustering.

    @dylanagreen
    dylanagreen committed Aug 31, 2020
    Configuration menu
    Copy the full SHA
    c4461da View commit details
    Browse the repository at this point in the history

Commits on Sep 1, 2020

  1. Improve performance on FOM_DETF training.

    @dylanagreen
    dylanagreen committed Sep 1, 2020
    Configuration menu
    Copy the full SHA
    89a386e View commit details
    Browse the repository at this point in the history

  2. Improvements to 8 bins (and beyond..?)

    @dylanagreen
    dylanagreen committed Sep 1, 2020
    Configuration menu
    Copy the full SHA
    79e73e5 View commit details
    Browse the repository at this point in the history

  3. Clean some docstrings in pca_cluster.py

    @dylanagreen
    dylanagreen committed Sep 1, 2020
    Configuration menu
    Copy the full SHA
    703b37c View commit details
    Browse the repository at this point in the history

  4. Cleanup binning_as_clustering. 

    - Additionally brings it in line with the new training methods.
    @dylanagreen
    dylanagreen committed Sep 1, 2020
    Configuration menu
    Copy the full SHA
    7d8772d View commit details
    Browse the repository at this point in the history

Commits on Sep 2, 2020

  1. Fix an embarrasing bug choosing learning rate.

    @dylanagreen
    dylanagreen committed Sep 2, 2020
    Configuration menu
    Copy the full SHA
    61c8028 View commit details
    Browse the repository at this point in the history

Commits on Sep 10, 2020

  1. Massively improve training performance. 

    - Use one-cycle policy rather than range test for training.
- Do not train over beta, approximate it as n_bins.
- Also speeds training to ~1/3 the time before.
    @dylanagreen
    dylanagreen committed Sep 10, 2020
    Configuration menu
    Copy the full SHA
    462716b View commit details
    Browse the repository at this point in the history

Commits on Sep 15, 2020

  1. Add optional buzzard parameter 

    - This parameter controls whether we should use the Buzzard learning
rate scheme for the FOM_DETF metric (and that metric only, as it is the
only one that shows improvement in this scheme).
    @dylanagreen
    dylanagreen committed Sep 15, 2020
    Configuration menu
    Copy the full SHA
    acfc3e4 View commit details
    Browse the repository at this point in the history

  2. Enable riz training. 

    - Also fixed a misspelling of the word "verbose".
    @dylanagreen
    dylanagreen committed Sep 15, 2020
    Configuration menu
    Copy the full SHA
    e1f330c View commit details
    Browse the repository at this point in the history