fixed up the computing conditional latents

tortoise/api.py

@@ -284,7 +284,7 @@ class TextToSpeech:
         if self.minor_optimizations:
             self.cvvp = self.cvvp.to(self.device)
 
-    def get_conditioning_latents(self, voice_samples, return_mels=False, verbose=False, progress=None, enforced_length=102400):
+    def get_conditioning_latents(self, voice_samples, return_mels=False, verbose=False, progress=None, enforced_length=None, chunk_tensors=False):
         """
         Transforms one or more voice_samples into a tuple (autoregressive_conditioning_latent, diffusion_conditioning_latent).
         These are expressive learned latents that encode aspects of the provided clips like voice, intonation, and acoustic
@@ -304,12 +304,30 @@ class TextToSpeech:
             
             diffusion_conds = []
             
-            for sample in tqdm_override(voice_samples, verbose=verbose, progress=progress, desc="Computing conditioning latents..."):
+            samples = [] # resample in its own pass to make things easier
+            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)
+                samples.append(torchaudio.functional.resample(sample, 22050, 24000))
-                chunks = torch.chunk(sample, int(sample.shape[-1] / enforced_length) + 1, dim=1)
 
-                for chunk in chunks:
+            if enforced_length is None:
+                for sample in tqdm_override(samples, verbose=verbose and len(samples) > 1, progress=progress if len(samples) > 1 else None, desc="Calculating size of best fit..."):
+                    if chunk_tensors:
+                        enforced_length = sample.shape[-1] if enforced_length is None else min( enforced_length, sample.shape[-1] )
+                    else:
+                        enforced_length = sample.shape[-1] if enforced_length is None else max( enforced_length, sample.shape[-1] )
+
+            print(f"Size of best fit: {enforced_length}")
+
+            chunks = []
+            if chunk_tensors:
+                for sample in tqdm_override(samples, verbose=verbose, progress=progress, desc="Slicing samples into chunks..."):
+                    sliced = torch.chunk(sample, int(sample.shape[-1] / enforced_length) + 1, dim=1)
+                    for s in sliced:
+                        chunks.append(s)
+            else:
+                chunks = samples
+                
+            for chunk in tqdm_override(chunks, verbose=verbose, progress=progress, desc="Computing conditioning latents..."):
                 chunk = pad_or_truncate(chunk, enforced_length)
                 cond_mel = wav_to_univnet_mel(chunk.to(self.device), do_normalization=False, device=self.device)
                 diffusion_conds.append(cond_mel)
@@ -424,6 +442,7 @@ class TextToSpeech:
         :return: Generated audio clip(s) as a torch tensor. Shape 1,S if k=1 else, (k,1,S) where S is the sample length.
                  Sample rate is 24kHz.
         """
+        self.diffusion.enable_fp16 = half_p
         deterministic_seed = self.deterministic_state(seed=use_deterministic_seed)
 
         text_tokens = torch.IntTensor(self.tokenizer.encode(text)).unsqueeze(0).to(self.device)
@@ -432,7 +451,7 @@ class TextToSpeech:
 
         auto_conds = None
         if voice_samples is not None:
-            auto_conditioning, diffusion_conditioning, auto_conds, _ = self.get_conditioning_latents(voice_samples, return_mels=True)
+            auto_conditioning, diffusion_conditioning, auto_conds, _ = self.get_conditioning_latents(voice_samples, return_mels=True, verbose=True)
         elif conditioning_latents is not None:
             auto_conditioning, diffusion_conditioning = conditioning_latents
         else:

tortoise/get_conditioning_latents.py

@@ -25,6 +25,6 @@ if __name__ == '__main__':
         for cond_path in cond_paths:
             c = load_audio(cond_path, 22050)
             conds.append(c)
-        conditioning_latents = tts.get_conditioning_latents(conds)
+        conditioning_latents = tts.get_conditioning_latents(conds, verbose=True)
         torch.save(conditioning_latents, os.path.join(args.output_path, f'{voice}.pth'))
 

tortoise/models/diffusion_decoder.py

@@ -141,7 +141,7 @@ class DiffusionTts(nn.Module):
             in_tokens=8193,
             out_channels=200,  # mean and variance
             dropout=0,
-            use_fp16=False,
+            use_fp16=True,
             num_heads=16,
             # Parameters for regularization.
             layer_drop=.1,