First result upload.

app.py

@@ -38,7 +38,7 @@ except Exception:
 
 MODEL_COLUMN_NAME = "model"
 
-column_names = {MODEL_COLUMN_NAME: "Model", "avg_wer": "Average WER ⬇️"}
+column_names = {MODEL_COLUMN_NAME: "Model", "avg_wer": "Average WER ⬇️", "avg_wil": "Average WIL"}
 
 leaderboard_df = pd.read_csv("benchmark.csv")  # TODO: Move to a different place, probably in the github
 leaderboard_df.loc[:, MODEL_COLUMN_NAME] = leaderboard_df[MODEL_COLUMN_NAME].apply(make_clickable_model)
@@ -76,4 +76,4 @@ with demo:
 scheduler = BackgroundScheduler()
 scheduler.add_job(restart_space, "interval", seconds=1800)
 scheduler.start()
-demo.queue(default_concurrency_limit=40).launch()
+demo.queue(default_concurrency_limit=40).launch()

benchmark.csv

@@ -1,3 +1,5 @@
-model,avg_wer,ivrit-ai/audio-labeled,ivrit-ai/audio-transcripts
-whisper-large-v2-d4,0.4,0.4,0.4
-whisper-v2-d3-e3,0.3,0.3,0.3
+model,avg_wer,avg_wil,ivrit-ai/eval-d1,ivrit-ai/saspeech,google/fleurs/he,mozilla-foundation/common_voice_17_0/he
+ivrit-ai/whisper-large-v2-d4,0.0,0.063,0.080,0.242,0.207
+ivrit-ai/whisper-v2-d3-e3,0.0,0.069,0.086,0.256,0.214
+openai/whisper-large-v2,0.0,0.083,0.100,0.276,0.234
+openai/whisper-large-v3,0.0,0.101,0.096,0.262,0.232