Update custom_interface_app.py
Browse files- custom_interface_app.py +6 -6
custom_interface_app.py
CHANGED
|
@@ -137,7 +137,7 @@ class ASR(Pretrained):
|
|
| 137 |
# Get audio length in seconds
|
| 138 |
audio_length = len(waveform) / sr
|
| 139 |
|
| 140 |
-
if audio_length >=
|
| 141 |
print(f"Audio is too long ({audio_length:.2f} seconds), splitting into segments")
|
| 142 |
# Detect non-silent segments
|
| 143 |
|
|
@@ -146,7 +146,7 @@ class ASR(Pretrained):
|
|
| 146 |
segments = []
|
| 147 |
current_segment = []
|
| 148 |
current_length = 0
|
| 149 |
-
max_duration =
|
| 150 |
|
| 151 |
|
| 152 |
for interval in non_silent_intervals:
|
|
@@ -204,7 +204,7 @@ class ASR(Pretrained):
|
|
| 204 |
# Get audio length in seconds
|
| 205 |
audio_length = len(waveform) / sr
|
| 206 |
|
| 207 |
-
if audio_length >=
|
| 208 |
print(f"Audio is too long ({audio_length:.2f} seconds), splitting into segments")
|
| 209 |
# Detect non-silent segments
|
| 210 |
|
|
@@ -213,7 +213,7 @@ class ASR(Pretrained):
|
|
| 213 |
segments = []
|
| 214 |
current_segment = []
|
| 215 |
current_length = 0
|
| 216 |
-
max_duration =
|
| 217 |
|
| 218 |
|
| 219 |
for interval in non_silent_intervals:
|
|
@@ -273,7 +273,7 @@ class ASR(Pretrained):
|
|
| 273 |
# Get audio length in seconds
|
| 274 |
audio_length = len(waveform) / sr
|
| 275 |
|
| 276 |
-
if audio_length >=
|
| 277 |
print(f"MMS Audio is too long ({audio_length:.2f} seconds), splitting into segments")
|
| 278 |
# Detect non-silent segments
|
| 279 |
non_silent_intervals = librosa.effects.split(waveform, top_db=20) # Adjust top_db for sensitivity
|
|
@@ -281,7 +281,7 @@ class ASR(Pretrained):
|
|
| 281 |
segments = []
|
| 282 |
current_segment = []
|
| 283 |
current_length = 0
|
| 284 |
-
max_duration =
|
| 285 |
|
| 286 |
|
| 287 |
for interval in non_silent_intervals:
|
|
|
|
| 137 |
# Get audio length in seconds
|
| 138 |
audio_length = len(waveform) / sr
|
| 139 |
|
| 140 |
+
if audio_length >= 20:
|
| 141 |
print(f"Audio is too long ({audio_length:.2f} seconds), splitting into segments")
|
| 142 |
# Detect non-silent segments
|
| 143 |
|
|
|
|
| 146 |
segments = []
|
| 147 |
current_segment = []
|
| 148 |
current_length = 0
|
| 149 |
+
max_duration = 20 * sr # Maximum segment duration in samples (20 seconds)
|
| 150 |
|
| 151 |
|
| 152 |
for interval in non_silent_intervals:
|
|
|
|
| 204 |
# Get audio length in seconds
|
| 205 |
audio_length = len(waveform) / sr
|
| 206 |
|
| 207 |
+
if audio_length >= 20:
|
| 208 |
print(f"Audio is too long ({audio_length:.2f} seconds), splitting into segments")
|
| 209 |
# Detect non-silent segments
|
| 210 |
|
|
|
|
| 213 |
segments = []
|
| 214 |
current_segment = []
|
| 215 |
current_length = 0
|
| 216 |
+
max_duration = 20 * sr # Maximum segment duration in samples (20 seconds)
|
| 217 |
|
| 218 |
|
| 219 |
for interval in non_silent_intervals:
|
|
|
|
| 273 |
# Get audio length in seconds
|
| 274 |
audio_length = len(waveform) / sr
|
| 275 |
|
| 276 |
+
if audio_length >= 20:
|
| 277 |
print(f"MMS Audio is too long ({audio_length:.2f} seconds), splitting into segments")
|
| 278 |
# Detect non-silent segments
|
| 279 |
non_silent_intervals = librosa.effects.split(waveform, top_db=20) # Adjust top_db for sensitivity
|
|
|
|
| 281 |
segments = []
|
| 282 |
current_segment = []
|
| 283 |
current_length = 0
|
| 284 |
+
max_duration = 20 * sr # Maximum segment duration in samples (20 seconds)
|
| 285 |
|
| 286 |
|
| 287 |
for interval in non_silent_intervals:
|