def synthesize(text): input = text + "|00-" + lang + "|" + lang # Change to Multi_TTS path sys.path.append( os.path.join(os.path.dirname(__file__), "dependencies/Multilingual_Text_to_Speech")) if "utils" in sys.modules: del sys.modules["utils"] from synthesize import synthesize from utils import build_model # Load Mulilingual pretrained model model = build_model( os.path.abspath("./dependencies/checkpoints/generated_switching.pyt")) model.eval() # generate spectogram spectogram = synthesize(model, "|" + input) # Change to WaveRNN Path sys.path.append( os.path.join(os.path.dirname(__file__), "dependencies/WaveRNN")) if "utils" in sys.modules: del sys.modules["utils"] from models.fatchord_version import WaveRNN from utils import hparams as hp from gen_wavernn import generate import torch # Load WaveRNN pretrained model hp.configure("hparams.py") model = WaveRNN( rnn_dims=hp.voc_rnn_dims, fc_dims=hp.voc_fc_dims, bits=hp.bits, pad=hp.voc_pad, upsample_factors=hp.voc_upsample_factors, feat_dims=hp.num_mels, compute_dims=hp.voc_compute_dims, res_out_dims=hp.voc_res_out_dims, res_blocks=hp.voc_res_blocks, hop_length=hp.hop_length, sample_rate=hp.sample_rate, mode=hp.voc_mode).to( torch.device('cuda' if torch.cuda.is_available() else 'cpu')) model.load( os.path.join(os.path.dirname(__file__), "dependencies/checkpoints/wavernn_weight.pyt")) waveform = generate(model, s, hp.voc_gen_batched, hp.voc_target, hp.voc_overlap) f = write("./temp/result.wav", "x") f.write(waveform) f.close()
sys.path.append(WAVERNN_FOLDER) from gen_wavernn import generate from utils import hparams as hp from models.fatchord_version import WaveRNN hp.configure(WAVERNN_FOLDER+'/hparams.py') model = WaveRNN(rnn_dims=hp.voc_rnn_dims, fc_dims=hp.voc_fc_dims, bits=hp.bits, pad=hp.voc_pad, upsample_factors=hp.voc_upsample_factors, feat_dims=hp.num_mels, compute_dims=hp.voc_compute_dims, res_out_dims=hp.voc_res_out_dims, res_blocks=hp.voc_res_blocks, hop_length=hp.hop_length, sample_rate=hp.sample_rate, mode=hp.voc_mode).to('cpu') model.load(CHECKPOINTS_FOLDER + "/" + wavernn_chpt) y = [] ix=1 while os.path.exists(CHR_FOLDER+"/"+str(ix)+".npy"): print("Found", CHR_FOLDER+"/"+str(ix)+".npy") y.append(np.load(CHR_FOLDER+"/"+str(ix)+".npy")) ix+=1 idx=1 for s in y: waveform = generate(model, s, hp.voc_gen_batched, hp.voc_target, hp.voc_overlap) sf.write("wg-"+str(idx)+".wav", waveform, hp.sample_rate) idx+=1
from gen_wavernn import generate from utils import hparams as hp from models.fatchord_version import WaveRNN hp.configure(WAVERNN_FOLDER+'/hparams.py') model = WaveRNN(rnn_dims=hp.voc_rnn_dims, fc_dims=hp.voc_fc_dims, bits=hp.bits, pad=hp.voc_pad, upsample_factors=hp.voc_upsample_factors, feat_dims=hp.num_mels, compute_dims=hp.voc_compute_dims, res_out_dims=hp.voc_res_out_dims, res_blocks=hp.voc_res_blocks, hop_length=hp.hop_length, sample_rate=hp.sample_rate, mode=hp.voc_mode).to('cuda') model.load(CHECKPOINTS_FOLDER + "/" + wavernn_chpt) y = [] y.append(np.load(TACOTRON_FOLDER + "/1.npy")) y.append(np.load(TACOTRON_FOLDER + "/2.npy")) y.append(np.load(TACOTRON_FOLDER + "/3.npy")) y.append(np.load(TACOTRON_FOLDER + "/4.npy")) y.append(np.load(TACOTRON_FOLDER + "/5.npy")) y.append(np.load(TACOTRON_FOLDER + "/6.npy")) y.append(np.load(TACOTRON_FOLDER + "/7.npy")) y.append(np.load(TACOTRON_FOLDER + "/8.npy")) waveforms = [generate(model, s, hp.voc_gen_batched, hp.voc_target, hp.voc_overlap) for s in y] for idx, w in enumerate(waveforms): sf.write("wg-"+str(idx+1)+".wav", w, hp.sample_rate)
model = WaveRNN(rnn_dims=hp.voc_rnn_dims, fc_dims=hp.voc_fc_dims, bits=hp.bits, pad=hp.voc_pad, upsample_factors=hp.voc_upsample_factors, feat_dims=hp.num_mels, compute_dims=hp.voc_compute_dims, res_out_dims=hp.voc_res_out_dims, res_blocks=hp.voc_res_blocks, hop_length=hp.hop_length, sample_rate=hp.sample_rate, mode=hp.voc_mode).to('cpu') model.load(CHECKPOINTS_FOLDER + "/" + wavernn_chpt) y = [] ix = 1 while os.path.exists(CHR_FOLDER + "/" + str(ix) + ".npy"): y.append(np.load(CHR_FOLDER + "/" + str(ix) + ".npy")) ix += 1 idx = 1 for s in y: waveform = generate(model, s, batched=True, target=11025, overlap=int(11025 / 4)) sf.write("wg-" + str(idx) + ".wav", waveform, hp.sample_rate) idx += 1
sys.path.append(WAVERNN_FOLDER) from gen_wavernn import generate from utils import hparams as hp from models.fatchord_version import WaveRNN hp.configure(WAVERNN_FOLDER + '/hparams.py') model = WaveRNN(rnn_dims=hp.voc_rnn_dims, fc_dims=hp.voc_fc_dims, bits=hp.bits, pad=hp.voc_pad, upsample_factors=hp.voc_upsample_factors, feat_dims=hp.num_mels, compute_dims=hp.voc_compute_dims, res_out_dims=hp.voc_res_out_dims, res_blocks=hp.voc_res_blocks, hop_length=hp.hop_length, sample_rate=hp.sample_rate, mode="RAW").to(device) model.load(CHECKPOINTS_FOLDER + "/" + wavernn_chpt) waveform = generate(model, np.load(cwd + "/tmp.npy"), batched=True, target=hp.voc_target, overlap=hp.voc_overlap) sf.write(cwd + "/tmp.wav", waveform, hp.sample_rate) sys.exit()