Commit History

  1. Add a button for "Exclude language modeling tasks from overall duel win score" Running

    743bbf3

  1. Another try with bokeh2iframe()

    24ab55a

  1. Try to make bokeh plots low-weight with bokeh2html()

    1cfab55

  1. Added cache self.leaderboard_scatter_plots

    bc1411a
  2. Added cache self.leaderboard_heatmaps

    d6b7a5a

  1. Heatmap: Added model_size_threshold constant

    e38b67a
  2. Heatmap: Fixed typo from singular to plural in axis labels

    6de4a42
  3. Heatmap: Rotate the graph from horizontal to vertical for easier reading and add the transpose parameter for easier switching in the future

    479d899

  1. Heatmap: First the larger models, then the smaller ones

    e84b5e6
  2. Heatmap: Aesthetic removal of gaps in axis labels

    7a318aa
  3. Heatmap: Change from abbreviations to category and task names

    f098bed
  4. Heatmap: Sort by category only when viewing all categories

    a07b5d1
  5. Heatmap: Second heatmap added

    d035edc
  6. Heatmap: Fix space typo

    d2d74b8
  7. Heatmap: If the size of the heatmap is not specified, it will be calculated

    5f8cc60
  8. Heatmap: Abbreviations for categories, created functions _prepare_tasks_categories_abbreviations() and get_leaderboard_heatmap(), and modification of create_heatmap()

    c2e1968

  1. Added the "precision" attribute to the leaderboard table

    c2d6748

  1. The "chat_template" attribute in the table is now a sign of its presence

    f09348d

  1. Truncate value of "chat_template" for table on web

    f83e2fa
  2. Interpretation of None changed from "null" to "none"

    7b8b773
  3. Changed order of attr. in LB table

    2e4d736
  4. Changed abbr. CAT → CHAT and explanation of abbr. ML

    a8986db

  1. Added more model attributes to the leaderboard table

    1222740

  1. Added selector of p_value

    cc83df6

  1. Added info message

    ffd19db
  2. Uploading corrected tournament_results

    b08adb9
  3. Optimization of the significance correction

    595ce2e
  4. Added tqdm to _correct_significance_in_tournament_results

    bfc0509
  5. Correcting significance in `tournament_results` with FDR correction method⁠

    ff10e2a

  1. Revert "Bokeh figure - changed Y axis label"⁠

    833a957
  2. Removed unnecessary calls of print()

    a6f409a
  3. Renamed `get_bokeh_figure` to `get_leaderboard_scatter_plot`

    182bc15
  4. Bokeh figure - changed Y axis label

    182ef2d
  5. Bokeh figure - change category according to the dropdown

    4d6c6ba

  1. Try analyze winscore with bokeh

    a2b6394

  1. Rename the submission date column from Release to Submitted

    5928b80

  1. Reduced processing in the critical section

    cfb07ff

  1. Tuning the duration of some notifications

    73f436c
  2. Upload only when it could be changed⁠

    e2d6d6c
  3. Do not read big JSON file twice

    0e8ff48

  1. More info notifications

    9d3d879
  2. Default sorting of leaderboard is by "average_score"⁠

    81fdab8
  3. For testing purpose improved fake tournament

    54b531c
  4. Locks only on confirmation of submission

    2e42a59

  1. Now only executes the duels that are missing⁠

    dfdc202

  1. Upload to the dataset the map from "submission_id" to "model_title"⁠

    8ae0f62

  1. Get task names only from "tasks_metadata.json"

    387e6ad
  2. Use specified metric for every task

    9225102

  1. Prepare to pandas 3.0

    c3d760a
  2. Repeat when connection to check significance server is timed out

    7d453b5