Update extract_feature_print.py
Browse files- extract_feature_print.py +14 -1
extract_feature_print.py
CHANGED
|
@@ -1,6 +1,6 @@
|
|
| 1 |
import os, sys, traceback
|
| 2 |
from transformers import HubertModel
|
| 3 |
-
import
|
| 4 |
from torch import nn
|
| 5 |
import torch
|
| 6 |
|
|
@@ -157,6 +157,19 @@ class Config:
|
|
| 157 |
)
|
| 158 |
config = Config()
|
| 159 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 160 |
#HuggingFacePlaceHolder = None
|
| 161 |
class HubertModelWithFinalProj(HubertModel):
|
| 162 |
def __init__(self, config):
|
|
|
|
| 1 |
import os, sys, traceback
|
| 2 |
from transformers import HubertModel
|
| 3 |
+
import librosa
|
| 4 |
from torch import nn
|
| 5 |
import torch
|
| 6 |
|
|
|
|
| 157 |
)
|
| 158 |
config = Config()
|
| 159 |
|
| 160 |
+
def load_audio(file, sample_rate):
|
| 161 |
+
try:
|
| 162 |
+
file = file.strip(" ").strip('"').strip("\n").strip('"').strip(" ")
|
| 163 |
+
audio, sr = sf.read(file)
|
| 164 |
+
if len(audio.shape) > 1:
|
| 165 |
+
audio = librosa.to_mono(audio.T)
|
| 166 |
+
if sr != sample_rate:
|
| 167 |
+
audio = librosa.resample(audio, orig_sr=sr, target_sr=sample_rate)
|
| 168 |
+
except Exception as error:
|
| 169 |
+
raise RuntimeError(f"An error occurred loading the audio: {error}")
|
| 170 |
+
|
| 171 |
+
return audio.flatten()
|
| 172 |
+
|
| 173 |
#HuggingFacePlaceHolder = None
|
| 174 |
class HubertModelWithFinalProj(HubertModel):
|
| 175 |
def __init__(self, config):
|