accept wav file only

AudioToText/condensedmodel.py

@@ -22,8 +22,6 @@ from tensorflow import keras
 from keras import layers
 import librosa
 import speech_recognition as sr
-import subprocess
-import os
 
 # MODEL LOSS
 def CTCLoss(y_true, y_pred):
@@ -135,9 +133,8 @@ def loadWeights():
 
     # Load CKPT to Model
     model.load_weights(output)
-s
 def load_wav(filename):
-    wav,_ = librosa.load(convertAudioToWav(filename), sr = 22050)
+    wav,_ = librosa.load(filename, sr = 22050)
 
     audio = tf.convert_to_tensor(
         wav,
@@ -205,7 +202,7 @@ def AudioToTextUsingModel(wav_file):
     return output_text
 
 def AudioToTextUsingAPI(audio_file):
-    AUDIO_FILE = load_wav(audio_file)
+    AUDIO_FILE = audio_file
     
     # use the audio file as the audio source
     

app.py

@@ -7,7 +7,7 @@ from AudioToText.condensedmodel import AudioToTextUsingAPI
 from AudioToText.condensedmodel import AudioToTextUsingModel
 
 
-st.title("Translation model test")
+st.title("FIRE COML Summer 2022 Translation Model")
 
 option = st.selectbox("Select input type:", ("Text input", "Audio input"))
 option2 = st.selectbox("Select translation language:", ("Spanish", "Chinese"))
@@ -25,9 +25,10 @@ if option == "Text input":
             st.write(translated)
             input_sentence = None
 else:
-    wav_sentence = st.file_uploader("Upload a .wav file:")
+    wav_sentence = st.file_uploader("Upload an audio file (.wav):", type=\
+        ["wav"])
     option3 = st.selectbox("Select audio to text model to use:", ("Our pretrained model", "Google API"))
-    if st.button("Submit .wav file"):
+    if st.button("Submit audio file"):
         if option3 == "Our pretrained model":
             input_list = AudioToTextUsingModel(wav_sentence)
             input_sentence = "".join(input_list)