change to cpu

utils.py

@@ -29,10 +29,13 @@ latent_map = {}
 
 def get_latents(chatbot_voice, xtts_model, voice_cleanup=False):
     global latent_map
-    if chatbot_voice not in latent_map:    
+    print("here")
+    if chatbot_voice not in latent_map:   
+        print("here1") 
         speaker_wav = f"examples/{chatbot_voice}.wav"
         if (voice_cleanup):
             try:
+                print("here2")
                 cleanup_filter="lowpass=8000,highpass=75,areverse,silenceremove=start_periods=1:start_silence=0:start_threshold=0.02,areverse,silenceremove=start_periods=1:start_silence=0:start_threshold=0.02" 
                 resample_filter="-ac 1 -ar 22050"
                 out_filename = speaker_wav + str(uuid.uuid4()) + ".wav"  #ffmpeg to know output format
@@ -45,6 +48,7 @@ def get_latents(chatbot_voice, xtts_model, voice_cleanup=False):
                 # There was an error - command exited with non-zero code
                 print("Error: failed filtering, use original microphone input")
         else:
+                print("here3")
                 speaker_wav=speaker_wav
         # gets condition latents from the model
         # returns tuple (gpt_cond_latent, speaker_embedding)