From cb7569ee5e580cc3570942befebc9d7d8dfdc9fa Mon Sep 17 00:00:00 2001 From: James Betker Date: Sat, 18 Jun 2022 10:40:48 -0600 Subject: [PATCH] resample real inputs for music_diffusion_fid --- codes/trainer/eval/music_diffusion_fid.py | 16 ++++++++++------ 1 file changed, 10 insertions(+), 6 deletions(-) diff --git a/codes/trainer/eval/music_diffusion_fid.py b/codes/trainer/eval/music_diffusion_fid.py index 25a12168..b41f7572 100644 --- a/codes/trainer/eval/music_diffusion_fid.py +++ b/codes/trainer/eval/music_diffusion_fid.py @@ -52,7 +52,7 @@ class MusicDiffusionFid(evaluator.Evaluator): model_var_type='learned_range', loss_type='mse', betas=get_named_beta_schedule('linear', 4000), conditioning_free=False, conditioning_free_k=1) self.dev = self.env['device'] - mode = opt_get(opt_eval, ['diffusion_type'], 'tts') + mode = opt_get(opt_eval, ['diffusion_type'], 'spec_decode') self.spec_decoder = get_mel2wav_model() self.projector = ContrastiveAudio(model_dim=512, transformer_heads=8, dropout=0, encoder_depth=8, mel_channels=256) @@ -141,7 +141,11 @@ class MusicDiffusionFid(evaluator.Evaluator): model_kwargs={'aligned_conditioning': gen_mel_denorm}) gen_wav = pixel_shuffle_1d(gen_wav, 16) - return gen_wav, real_resampled, gen_mel, mel_norm, sample_rate + real_wav = self.spectral_diffuser.p_sample_loop(self.spec_decoder, output_shape, + model_kwargs={'aligned_conditioning': mel}) + real_wav = pixel_shuffle_1d(real_wav, 16) + + return gen_wav, real_wav.squeeze(0), gen_mel, mel_norm, sample_rate def perform_partial_diffusion_from_codes_quant(self, audio, sample_rate=22050, partial_low=0, partial_high=256): if sample_rate != sample_rate: @@ -271,15 +275,15 @@ class MusicDiffusionFid(evaluator.Evaluator): if __name__ == '__main__': diffusion = load_model_from_config('X:\\dlas\\experiments\\train_music_diffusion_tfd_quant.yml', 'generator', also_load_savepoint=False, - load_path='X:\\dlas\\experiments\\train_music_diffusion_tfd11\\models\\24000_generator_ema.pth' + load_path='X:\\dlas\\experiments\\train_music_diffusion_tfd12\\models\\41500_generator_ema.pth' ).cuda() opt_eval = {'path': 'Y:\\split\\yt-music-eval', # eval music, mostly electronica. :) #'path': 'E:\\music_eval', # this is music from the training dataset, including a lot more variety. 'diffusion_steps': 200, - 'conditioning_free': False, 'conditioning_free_k': 1, - 'diffusion_schedule': 'cosine', 'diffusion_type': 'from_codes_quant', + 'conditioning_free': True, 'conditioning_free_k': 2, + 'diffusion_schedule': 'linear', 'diffusion_type': 'from_codes_quant', #'partial_low': 128, 'partial_high': 192 } - env = {'rank': 0, 'base_path': 'D:\\tmp\\test_eval_music', 'step': 600, 'device': 'cuda', 'opt': {}} + env = {'rank': 0, 'base_path': 'D:\\tmp\\test_eval_music', 'step': 605, 'device': 'cuda', 'opt': {}} eval = MusicDiffusionFid(diffusion, opt_eval, env) print(eval.perform_eval())