Upload 2 files

config.yml

@@ -0,0 +1,56 @@
+log_dir: "Models/run_redecoder_wavenet"
+save_freq: 1
+log_interval: 10
+save_interval: 1000
+device: "cuda"
+epochs: 1000 # number of epochs for first stage training (pre-training)
+batch_size: 4
+batch_length: 100 # maximum duration of audio in a batch (in seconds)
+max_len: 80 # maximum number of frames
+pretrained_model: ""
+pretrained_encoder: "./temp_ckpt.pth"
+load_only_params: False # set to true if do not want to load epoch numbers and optimizer parameters
+
+F0_path: "modules/JDC/bst.t7"
+
+data_params:
+  train_data: "./data/train.txt"
+  val_data: "./data/val.txt"
+  root_path: "./data/"
+
+preprocess_params:
+  sr: 24000
+  spect_params:
+    n_fft: 2048
+    win_length: 1200
+    hop_length: 300
+
+model_params:
+  encoder_causal: True
+  decoder_causal: False
+  encoder_lstm: 2
+  decoder_lstm: 0
+  n_c_codebooks: 2
+  n_p_codebooks: 1
+  timbre_norm: True
+  separate_prosody_encoder: True
+  encoder_type: 'wavenet' # should be one of mamba, transformer or wavenet
+  wavenet_embed_dim: 512
+  mamba_embed_dim: 768
+  prob_random_mask_prosody: 1.0
+  prob_random_mask_content: [0.0, 1.0]
+
+  DAC:
+    encoder_dim: 64
+    encoder_rates: [2, 5, 5, 6]
+    decoder_dim: 1536
+    decoder_rates: [ 6, 5, 5, 2 ]
+    sr: 24000
+
+loss_params:
+  base_lr: 0.0001
+  discriminator_iter_start: 2000
+  lambda_spk: 1.0
+  lambda_mel: 45
+  lambda_f0: 1.0
+  lambda_uv: 1.0

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c37f88e294cea2879a816805a400ef938d532732fb3fec7533521790e7b289b1
+size 637487702