DontPlanToEnd commited on
Commit
e913f3c
β€’
1 Parent(s): fb6ea92

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +2 -2
app.py CHANGED
@@ -295,7 +295,7 @@ with GraInter:
295
  *Because this leaderboard is just based on one short story generation, it obviously isn't going to be perfect*
296
  """)
297
 
298
- with gr.TabItem("Anime Rating Prediction"):
299
  leaderboard_df_arp = leaderboard_df.sort_values(by='Score πŸ†', ascending=False)
300
  leaderboard_df_arp_na = leaderboard_df_arp[leaderboard_df_arp[['Dif', 'Cor']].isna().any(axis=1)]
301
  leaderboard_df_arp = leaderboard_df_arp[~leaderboard_df_arp[['Dif', 'Cor']].isna().any(axis=1)]
@@ -313,7 +313,7 @@ with GraInter:
313
  gr.Markdown("""
314
  *This is a leaderboard of one of the questions from the UGI-Leaderboard. It doesn't use the decensoring system prompt the other questions do.*
315
  <br><br>
316
- **Anime Rating Prediction Leaderboard:** This leaderboard is meant to be a way to measure a model's ability to give intelligent recommendations. Given a user's list of ~300 anime ratings (1-10), the model is then given a different (and shorter) list of anime and is tasked with estimating what the user will rate each of them.
317
  <br>
318
  **Dif:** The average difference between the predicted and actual ratings of each anime.
319
  <br>
 
295
  *Because this leaderboard is just based on one short story generation, it obviously isn't going to be perfect*
296
  """)
297
 
298
+ with gr.TabItem("Rating Prediction"):
299
  leaderboard_df_arp = leaderboard_df.sort_values(by='Score πŸ†', ascending=False)
300
  leaderboard_df_arp_na = leaderboard_df_arp[leaderboard_df_arp[['Dif', 'Cor']].isna().any(axis=1)]
301
  leaderboard_df_arp = leaderboard_df_arp[~leaderboard_df_arp[['Dif', 'Cor']].isna().any(axis=1)]
 
313
  gr.Markdown("""
314
  *This is a leaderboard of one of the questions from the UGI-Leaderboard. It doesn't use the decensoring system prompt the other questions do.*
315
  <br><br>
316
+ **Rating Prediction Leaderboard:** This leaderboard is meant to be a way to measure a model's ability to give intelligent recommendations. Given a user's list of ~300 anime ratings (1-10), the model is then given a different (and shorter) list of anime and is tasked with estimating what the user will rate each of them.
317
  <br>
318
  **Dif:** The average difference between the predicted and actual ratings of each anime.
319
  <br>