added flag (--cond-latent-max-chunk-size) that should restrict the maximum chunk size when chunking for calculating conditional latents, to avoid OOMing on VRAM

app.py

@@ -27,7 +27,7 @@ def generate(text, delimiter, emotion, prompt, voice, mic_audio, preset, seed, c
     
     if voice_samples is not None:
         sample_voice = voice_samples[0]
-        conditioning_latents = tts.get_conditioning_latents(voice_samples, progress=progress)
+        conditioning_latents = tts.get_conditioning_latents(voice_samples, progress=progress, max_chunk_size=args.cond_latent_max_chunk_size)
         torch.save(conditioning_latents, os.path.join(f'./tortoise/voices/{voice}/', f'cond_latents.pth'))
         voice_samples = None
     else:
@@ -265,6 +265,7 @@ if __name__ == "__main__":
     parser = argparse.ArgumentParser()
     parser.add_argument("--share", action='store_true', help="Lets Gradio return a public URL to use anywhere")
     parser.add_argument("--low-vram", action='store_true', help="Disables some optimizations that increases VRAM usage")
+    parser.add_argument("--cond-latent-max-chunk-size", type=int, default=None, help="Sets an upper limit to audio chunk size when computing conditioning latents")
     args = parser.parse_args()
 
     tts = TextToSpeech(minor_optimizations=not args.low_vram)

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=None, chunk_tensors=False):
+    def get_conditioning_latents(self, voice_samples, return_mels=False, verbose=False, progress=None, chunk_size=None, max_chunk_size=None, chunk_tensors=True):
         """
         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
@@ -309,26 +309,29 @@ class TextToSpeech:
                 # The diffuser operates at a sample rate of 24000 (except for the latent inputs)
                 samples.append(torchaudio.functional.resample(sample, 22050, 24000))
 
-            if enforced_length is None:
+            if chunk_size 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] )
+                        chunk_size = sample.shape[-1] if chunk_size is None else min( chunk_size, sample.shape[-1] )
                     else:
-                        enforced_length = sample.shape[-1] if enforced_length is None else max( enforced_length, sample.shape[-1] )
+                        chunk_size = sample.shape[-1] if chunk_size is None else max( chunk_size, sample.shape[-1] )
 
-            print(f"Size of best fit: {enforced_length}")
+            print(f"Size of best fit: {chunk_size}")
+            if max_chunk_size is not None and chunk_size > max_chunk_size:
+                chunk_size = max_chunk_size
+                print(f"Chunk size exceeded, clamping to: {max_chunk_size}")
 
             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)
+                    sliced = torch.chunk(sample, int(sample.shape[-1] / chunk_size) + 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)
+                chunk = pad_or_truncate(chunk, chunk_size)
                 cond_mel = wav_to_univnet_mel(chunk.to(self.device), do_normalization=False, device=self.device)
                 diffusion_conds.append(cond_mel)
 
@@ -460,7 +463,7 @@ class TextToSpeech:
         diffusion_conditioning = diffusion_conditioning.to(self.device)
 
         diffuser = load_discrete_vocoder_diffuser(desired_diffusion_steps=diffusion_iterations, cond_free=cond_free, cond_free_k=cond_free_k)
-        
+
         with torch.no_grad():
             samples = []
             num_batches = num_autoregressive_samples // self.autoregressive_batch_size