Update app.py

app.py

@@ -1,20 +1,20 @@
-from nemo.collections.asr.models import NeuralDiarizer
+# from nemo.collections.asr.models import NeuralDiarizer
 import gradio as gr
 import pandas as pd
 
 device = "cuda" if torch.cuda.is_available() else "cpu"
 
-model = NeuralDiarizer.from_pretrained("diar_msdd_telephonic").to(device)
+# model = NeuralDiarizer.from_pretrained("diar_msdd_telephonic").to(device)
 
 def run_diarization(path1):
-    annotation = model(path1)
-    rttm=annotation.to_rttm()
+    # annotation = model(path1)
+    # rttm=annotation.to_rttm()
     df = pd.DataFrame(columns=['start_time', 'end_time', 'speaker'])
-    for idx,line in enumerate(rttm.splitlines()):
-        split = line.split()
-        start_time, duration, speaker = split[3], split[4], split[7]
-        end_time = float(start_time) + float(duration)
-        df.loc[idx] = start_time, end_time, speaker
+    # for idx,line in enumerate(rttm.splitlines()):
+    #     split = line.split()
+    #     start_time, duration, speaker = split[3], split[4], split[7]
+    #     end_time = float(start_time) + float(duration)
+    #     df.loc[idx] = start_time, end_time, speaker
     return df
 	
 inputs = [
@@ -29,8 +29,8 @@ description = (
 
 article = (
     "<p style='text-align: center'>"
-    "<a href='https://huggingface.co/nvidia/speakerverification_en_titanet_large' target='_blank'>🎙️ Learn more about TitaNet model</a> | "
-    "<a href='https://arxiv.org/pdf/2110.04410.pdf' target='_blank'>📚 TitaNet paper</a> | "
+    "<a href='https://catalog.ngc.nvidia.com/orgs/nvidia/teams/nemo/models/diar_msdd_telephonic' target='_blank'>🎙️ Learn more about MSDD model</a> | "
+    "<a href='https://arxiv.org/abs/2203.15974' target='_blank'>📚 MSDD paper</a> | "
     "<a href='https://github.com/NVIDIA/NeMo' target='_blank'>🧑‍💻 Repository</a>"
     "</p>"
 )