dmet/ai-voice-cloning
1
0
Fork 0
forked from mrq/ai-voice-cloning
Code Pull Requests Activity

rude

Browse Source
This commit is contained in:
mrq 2023-02-18 17:23:44 +00:00
parent 13d466baf5
commit 1f4cdcb8a9
1 changed files with 4 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
src/utils.py
View File

@ -172,7 +172,7 @@ def generate(
progress(0, desc="Loading voice...") progress(0, desc="Loading voice...")
voice_samples, conditioning_latents = load_voice(voice) voice_samples, conditioning_latents = load_voice(voice)
if voice_samples is not None and len(voice_samples) > 0: if voice_samples and len(voice_samples) > 0:
sample_voice = torch.cat(voice_samples, dim=-1).squeeze().cpu() sample_voice = torch.cat(voice_samples, dim=-1).squeeze().cpu()
conditioning_latents = tts.get_conditioning_latents(voice_samples, return_mels=not args.latents_lean_and_mean, progress=progress, slices=voice_latents_chunks, force_cpu=args.force_cpu_for_conditioning_latents) conditioning_latents = tts.get_conditioning_latents(voice_samples, return_mels=not args.latents_lean_and_mean, progress=progress, slices=voice_latents_chunks, force_cpu=args.force_cpu_for_conditioning_latents)
@ -185,7 +185,8 @@ def generate(
else: else:
if conditioning_latents is not None: if conditioning_latents is not None:
sample_voice, _ = load_voice(voice, load_latents=False) sample_voice, _ = load_voice(voice, load_latents=False)
sample_voice = torch.cat(sample_voice, dim=-1).squeeze().cpu() if sample_voice and len(sample_voice) > 0:
sample_voice = torch.cat(sample_voice, dim=-1).squeeze().cpu()
else: else:
sample_voice = None sample_voice = None
@ -392,7 +393,7 @@ def generate(
fixed_output_voices.append(fixed) fixed_output_voices.append(fixed)
output_voices = fixed_output_voices output_voices = fixed_output_voices
if voice is not None and conditioning_latents is not None: if voice and voice != "random" and conditioning_latents is not None:
with open(f'{get_voice_dir()}/{voice}/cond_latents.pth', 'rb') as f: with open(f'{get_voice_dir()}/{voice}/cond_latents.pth', 'rb') as f:
info['latents'] = base64.b64encode(f.read()).decode("ascii") info['latents'] = base64.b64encode(f.read()).decode("ascii")