esc-benchmark -> esc-bench

app.py

@@ -105,7 +105,7 @@ st.markdown(
    """
    ## Submitting to ESC
    \n
-   To submit to ESC, download the audio data for the nine mandatory ESC test sets from [esc-datasets](https://huggingface.co/datasets/esc-benchmark/esc-datasets). The test sets contain audio data only. Evaluate your system on the nine test sets by generating predictions for the unlabelled audio samples. For each test set, save the predictions to a .txt file in the order that the audio samples are provided, with one prediction per line. Name the .txt file according to the ESC test set names shown in the table (e.g. the predictions for LibriSpeech test-clean should be named ls-clean.txt).
+   To submit to ESC, download the audio data for the nine mandatory ESC test sets from [esc-datasets](https://huggingface.co/datasets/esc-bench/esc-datasets). The test sets contain audio data only. Evaluate your system on the nine test sets by generating predictions for the unlabelled audio samples. For each test set, save the predictions to a .txt file in the order that the audio samples are provided, with one prediction per line. Name the .txt file according to the ESC test set names shown in the table (e.g. the predictions for LibriSpeech test-clean should be named ls-clean.txt).
    \n
    Once you have evaluated your system on all nine test sets, move the predictions into one folder and zip it. The name you assign to the zipped folder will be the name that is shown on the table (e.g. whisper-aed.zip will be displayed as whisper-aed). Upload your zipped submissions for scoring and placement on the leaderboard.
 """