Add use_gpt_tts script

codes/scripts/audio/gen/use_gpt_tts.py

@@ -0,0 +1,81 @@
+import argparse
+import os
+import random
+
+import torch
+import torchaudio
+import yaml
+
+from data.audio.unsupervised_audio_dataset import load_audio
+from data.util import is_audio_file, find_files_of_type
+from models.tacotron2.text import text_to_sequence
+from scripts.audio.gen.speech_synthesis_utils import do_spectrogram_diffusion, \
+    load_discrete_vocoder_diffuser, wav_to_mel, convert_mel_to_codes
+from trainer.injectors.base_injectors import MelSpectrogramInjector
+from utils.audio import plot_spectrogram
+from utils.options import Loader
+from utils.util import load_model_from_config
+import torch.nn.functional as F
+
+
+def do_vocoding(dvae, vocoder, diffuser, codes, cond=None, plot_spec=False):
+    return
+
+
+def load_conditioning_candidates(path, num_conds, sample_rate=22050, cond_length=44100):
+    candidates = find_files_of_type('img', path, qualifier=is_audio_file)[0]
+    # Sample with replacement. This can get repeats, but more conveniently handles situations where there are not enough candidates.
+    related_mels = []
+    for k in range(num_conds):
+        rel_clip = load_audio(candidates[k], sample_rate)
+        gap = rel_clip.shape[-1] - cond_length
+        if gap < 0:
+            rel_clip = F.pad(rel_clip, pad=(0, abs(gap)))
+        elif gap > 0:
+            rand_start = random.randint(0, gap)
+            rel_clip = rel_clip[:, rand_start:rand_start + cond_length]
+        mel_clip = MelSpectrogramInjector({'in': 'wav', 'out': 'mel'},{})({'wav': rel_clip.unsqueeze(0)})['mel'].squeeze(0)
+        related_mels.append(mel_clip)
+    return torch.stack(related_mels, dim=0)
+
+
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser()
+    parser.add_argument('-opt_diffuse', type=str, help='Path to options YAML file used to train the diffusion model', default='X:\\dlas\\experiments\\train_diffusion_vocoder_with_cond_new_dvae.yml')
+    parser.add_argument('-diffusion_model_name', type=str, help='Name of the diffusion model in opt.', default='generator')
+    parser.add_argument('-diffusion_model_path', type=str, help='Diffusion model checkpoint to load.', default='X:\\dlas\\experiments\\train_diffusion_vocoder_with_cond_new_dvae_full\\models\\6100_generator_ema.pth')
+    parser.add_argument('-dvae_model_name', type=str, help='Name of the DVAE model in opt.', default='dvae')
+    parser.add_argument('-opt_gpt_tts', type=str, help='Path to options YAML file used to train the GPT-TTS model', default='X:\\dlas\\experiments\\train_gpt_tts.yml')
+    parser.add_argument('-gpt_tts_model_name', type=str, help='Name of the GPT TTS model in opt.', default='gpt')
+    parser.add_argument('-gpt_tts_model_path', type=str, help='GPT TTS model checkpoint to load.', default='X:\\dlas\\experiments\\train_gpt_tts\\models\\22000_gpt.pth')
+    parser.add_argument('-text', type=str, help='Text to speak.', default="I'm a language model that has learned to speak.")
+    parser.add_argument('-cond_path', type=str, help='Folder containing conditioning samples.', default='Z:\\clips\\books1\\3042_18_Holden__000000000')
+    parser.add_argument('-num_cond', type=int, help='Number of conditioning samples to load.', default=3)
+    args = parser.parse_args()
+
+    print("Loading GPT TTS..")
+    with open(args.opt_gpt_tts, mode='r') as f:
+        gpt_opt = yaml.load(f, Loader=Loader)
+    gpt_opt['networks'][args.gpt_tts_model_name]['kwargs']['checkpointing'] = False  # Required for beam search
+    gpt = load_model_from_config(preloaded_options=gpt_opt, model_name=args.gpt_tts_model_name, also_load_savepoint=False, load_path=args.gpt_tts_model_path)
+
+    print("Loading data..")
+    text = torch.IntTensor(text_to_sequence(args.text, ['english_cleaners'])).unsqueeze(0).cuda()
+    conds = load_conditioning_candidates(args.cond_path, args.num_cond).unsqueeze(0).cuda()
+
+    print("Performing GPT inference..")
+    codes = gpt.inference(text, conds, num_beams=4)  #TODO: check the text length during training and match that during inference.
+
+    # Delete the GPT TTS model to free up GPU memory
+    del gpt
+
+    print("Loading DVAE..")
+    dvae = load_model_from_config(args.opt_diffuse, args.dvae_model_name)
+    print("Loading Diffusion Model..")
+    diffusion = load_model_from_config(args.opt_diffuse, args.diffusion_model_name, also_load_savepoint=False, load_path=args.diffusion_model_path)
+    diffuser = load_discrete_vocoder_diffuser()
+
+    print("Performing vocoding..")
+    wav = do_spectrogram_diffusion(diffusion, dvae, diffuser, codes, conds[:, 0], spectrogram_compression_factor=128, plt_spec=True)
+    torchaudio.save('gpt_tts_output.wav', wav.squeeze(0), 10025)

codes/utils/util.py

@@ -467,9 +467,12 @@ def clip_grad_norm(parameters: list, parameter_names: list, max_norm: float, nor
 
 
 Loader, Dumper = OrderedYaml()
-def load_model_from_config(cfg_file, model_name=None, dev='cuda', also_load_savepoint=True, load_path=None):
-    with open(cfg_file, mode='r') as f:
-        opt = yaml.load(f, Loader=Loader)
+def load_model_from_config(cfg_file=None, model_name=None, dev='cuda', also_load_savepoint=True, load_path=None, preloaded_options=None):
+    if preloaded_options is not None:
+        opt = preloaded_options
+    else:
+        with open(cfg_file, mode='r') as f:
+            opt = yaml.load(f, Loader=Loader)
     if model_name is None:
         model_cfg = opt['networks'].values()
         model_name = next(opt['networks'].keys())