update

inference/base_tts_infer.py

@@ -78,7 +78,7 @@ class BaseTTSInfer:
         # processed ref audio
         ref_audio = inp['ref_audio']
         processed_ref_audio = 'example/temp.wav'
-        voice_encoder = VoiceEncoder().cuda()
+        voice_encoder = VoiceEncoder().to(self.device)
         encoder = [self.ph_encoder, self.word_encoder]
         EmotionEncoder.load_model(self.hparams['emotion_encoder_path'])
         binarizer_cls = self.hparams.get("binarizer_cls", 'data_gen.tts.base_binarizerr.BaseBinarizer')
@@ -185,7 +185,7 @@ class BaseTTSInfer:
         input_values = self.asr_processor(audio_input, sampling_rate=sample_rate, return_tensors="pt").input_values
 
         # retrieve logits & take argmax
-        logits = self.asr_model(input_values.cuda()).logits
+        logits = self.asr_model(input_values.to(self.device)).logits
         predicted_ids = torch.argmax(logits, dim=-1)
 
         # transcribe