diff --git a/tortoise/api.py b/tortoise/api.py
index 57d01594..39d029a9 100644
--- a/tortoise/api.py
+++ b/tortoise/api.py
@@ -208,7 +208,7 @@ class TextToSpeech:
 
     def __init__(self, autoregressive_batch_size=None, models_dir=MODELS_DIR, 
                  enable_redaction=True, kv_cache=False, use_deepspeed=False, half=False, device=None,
-                 tokenizer_vocab_file=None, tokenizer_basic=False):
+                 tokenizer_vocab_file=None, tokenizer_basic=False, device_only=False):
 
         """
         Constructor
@@ -263,16 +263,27 @@ def __init__(self, autoregressive_batch_size=None, models_dir=MODELS_DIR,
         self.vocoder.load_state_dict(torch.load(get_model_path('vocoder.pth', models_dir), map_location=torch.device('cpu'))['model_g'])
         self.vocoder.eval(inference=True)
 
+        self.device_only = device_only
+
+        if self.device_only:
+            self.autoregressive = self.autoregressive.to(self.device)
+            self.diffusion = self.diffusion.to(self.device)
+            self.clvp = self.clvp.to(self.device)
+            self.vocoder = self.vocoder.to(self.device)
+
         # Random latent generators (RLGs) are loaded lazily.
         self.rlg_auto = None
         self.rlg_diffusion = None
+
     @contextmanager
     def temporary_cuda(self, model):
-        m = model.to(self.device)
-        yield m
-        m = model.cpu()
+        if not self.device_only:
+            m = model.to(self.device)
+            yield m
+            m = model.cpu()
+        else:
+            yield model
 
-    
     def load_cvvp(self):
         """Load CVVP model."""
         self.cvvp = CVVP(model_dim=512, transformer_heads=8, dropout=0, mel_codes=8192, conditioning_enc_depth=8, cond_mask_percentage=0,
@@ -295,22 +306,31 @@ def get_conditioning_latents(self, voice_samples, return_mels=False):
             for vs in voice_samples:
                 auto_conds.append(format_conditioning(vs, device=self.device))
             auto_conds = torch.stack(auto_conds, dim=1)
-            self.autoregressive = self.autoregressive.to(self.device)
-            auto_latent = self.autoregressive.get_conditioning(auto_conds)
-            self.autoregressive = self.autoregressive.cpu()
+
+            if not self.device_only:
+                self.autoregressive = self.autoregressive.to(self.device)
+                auto_latent = self.autoregressive.get_conditioning(auto_conds)
+                self.autoregressive = self.autoregressive.cpu()
+            else:
+                auto_latent = self.autoregressive.get_conditioning(auto_conds)
 
             diffusion_conds = []
+
             for sample in voice_samples:
                 # The diffuser operates at a sample rate of 24000 (except for the latent inputs)
                 sample = torchaudio.functional.resample(sample, 22050, 24000)
                 sample = pad_or_truncate(sample, 102400)
                 cond_mel = wav_to_univnet_mel(sample.to(self.device), do_normalization=False, device=self.device)
                 diffusion_conds.append(cond_mel)
+                
             diffusion_conds = torch.stack(diffusion_conds, dim=1)
 
-            self.diffusion = self.diffusion.to(self.device)
-            diffusion_latent = self.diffusion.get_conditioning(diffusion_conds)
-            self.diffusion = self.diffusion.cpu()
+            if not self.device_only:
+                self.diffusion = self.diffusion.to(self.device)
+                diffusion_latent = self.diffusion.get_conditioning(diffusion_conds)
+                self.diffusion = self.diffusion.cpu()
+            else:
+                diffusion_latent = self.diffusion.get_conditioning(diffusion_conds)
 
         if return_mels:
             return auto_latent, diffusion_latent, auto_conds, diffusion_conds