full run

src/predict.py

@@ -153,12 +153,8 @@ class Predictor:
         bucket_name = 'demovidelyuseruploads'
         if b'WAVE' in header:
             file_format = 'wav'
-        elif header.startswith((b'\xff\xfb', b'\xff\xf3', b'\xff\xe3', b'\xff\xfa')):
-            file_format = 'mp3'
         else:
-            file_format = 'unknown'
-        if file_format == 'unknown':
-            return {'error':'unrecognized file format, encode audio file as base64 str'}
+            file_format = 'mp3'
 
         unique_filename = f"{uuid.uuid4()}"
 
@@ -173,6 +169,8 @@ class Predictor:
             os.remove(local_filename)
         print(wav_filename)
 
+        model,sampler = self.model,self.sampler
+        result = self.process_audio_file(wav_filename,'this is a sample test to quickly run this model and resample',model,sampler)
         # if cut_audio > 0, means it was set
         if cut_audio > 0:
             #need to cut
@@ -323,6 +321,8 @@ class Predictor:
         bucket_name = 'demovidelyuseruploads'
         local_file_path = os.path.join(raw_dir,s3_key)
         self.download_file_from_s3(self.s3_client,bucket_name,s3_key,local_file_path)
+        model,sampler = self.model,self.sampler
+        result = self.process_audio_file(local_file_path,'this is a sample test to quickly run this model and resample',model,sampler)
 
         _,tone_color_converter = self.base_speaker_tts,self.tone_color_converter
         reference_speaker = local_file_path