NickVerri commited on
Commit
1e58243
·
verified ·
1 Parent(s): ca338e2

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +12 -4
app.py CHANGED
@@ -131,16 +131,18 @@ if uploaded_file:
131
 
132
  # 1. Diarization
133
  try:
134
- # Fix for pyannote.audio 3.0+: 'use_auth_token' is now just 'token'
 
135
  try:
136
  pipeline = Pipeline.from_pretrained(
137
- "pyannote/speaker-diarization@2.1",
 
138
  token=ACTIVE_HF_TOKEN
139
  )
140
  except TypeError:
141
- # Fallback for older versions of pyannote.audio
142
  pipeline = Pipeline.from_pretrained(
143
- "pyannote/speaker-diarization@2.1",
144
  use_auth_token=ACTIVE_HF_TOKEN
145
  )
146
 
@@ -179,6 +181,12 @@ if uploaded_file:
179
 
180
  if "transcript" in st.session_state:
181
  st.divider()
 
 
 
 
 
 
182
  brief = st.text_area("Creative Brief", placeholder="e.g. Focus on the yeast story, remove the interviewer.")
183
 
184
  if st.button("Step 2: Create EDL"):
 
131
 
132
  # 1. Diarization
133
  try:
134
+ # Fixed for recent pyannote.audio: separate model and revision
135
+ model_id = "pyannote/speaker-diarization"
136
  try:
137
  pipeline = Pipeline.from_pretrained(
138
+ model_id,
139
+ revision="2.1",
140
  token=ACTIVE_HF_TOKEN
141
  )
142
  except TypeError:
143
+ # Fallback if 'revision' isn't supported in old versions
144
  pipeline = Pipeline.from_pretrained(
145
+ f"{model_id}@2.1",
146
  use_auth_token=ACTIVE_HF_TOKEN
147
  )
148
 
 
181
 
182
  if "transcript" in st.session_state:
183
  st.divider()
184
+
185
+ # Diarization Preview to help the user identify roles
186
+ with st.expander("Transcript Preview (Diarized)"):
187
+ for seg in st.session_state.transcript[:20]: # Show first 20 segments
188
+ st.write(f"**{seg['speaker']}:** {seg['text']}")
189
+
190
  brief = st.text_area("Creative Brief", placeholder="e.g. Focus on the yeast story, remove the interviewer.")
191
 
192
  if st.button("Step 2: Create EDL"):