update

configs/combsub.yaml

@@ -5,12 +5,12 @@ data:
   sampling_rate: 44100
   block_size: 512 # Equal to hop_length
   duration: 2 # Audio duration during training, must be less than the duration of the shortest audio clip
-  encoder: 'hubertsoft' # 'hubertsoft', 'hubertbase', 'hubertbase768', 'contentvec', 'contentvec768' or 'contentvec768l12' or 'cnhubertsoftfish'
+  encoder: 'contentvec' # 'hubertsoft', 'hubertbase', 'hubertbase768', 'contentvec', 'contentvec768' or 'contentvec768l12' or 'cnhubertsoftfish'
   cnhubertsoft_gate: 10 # only use with cnhubertsoftfish
   encoder_sample_rate: 16000
   encoder_hop_size: 320
   encoder_out_channels: 256 # 768 if using 'hubertbase768', 'contentvec768' or 'contentvec768l12' encoder
-  encoder_ckpt: pretrain/hubert/hubert-soft-0d54a1f4.pt
+  encoder_ckpt: /workspace/DDSP-SVC/pretrain/hubert/checkpoint_best_legacy_500.pt
   train_path: data/train # Create a folder named "audio" under this path and put the audio clip in it
   valid_path: data/val # Create a folder named "audio" under this path and put the audio clip in it
 model:
@@ -18,7 +18,7 @@ model:
   n_spk: 1 # max number of different speakers
 enhancer:
     type: 'nsf-hifigan'
-    ckpt: 'pretrain/nsf_hifigan/model'
+    ckpt: '/workspace/DDSP-SVC/pretrain/nsf_hifigan/model'
 loss:
   fft_min: 256
   fft_max: 2048
@@ -28,10 +28,10 @@ env:
   expdir: exp/combsub-test
   gpu_id: 0
 train:
-  num_workers: 2 # If your cpu and gpu are both very strong, set to 0 may be faster!
-  batch_size: 24
+  num_workers: 0 # If your cpu and gpu are both very strong, set to 0 may be faster!
+  batch_size: 128
   cache_all_data: true # Save Internal-Memory or Graphics-Memory if it is false, but may be slow
-  cache_device: 'cpu' # Set to 'cuda' to cache the data into the Graphics-Memory, fastest speed for strong gpu
+  cache_device: 'cuda' # Set to 'cuda' to cache the data into the Graphics-Memory, fastest speed for strong gpu
   cache_fp16: true
   epochs: 100000
   interval_log: 10

configs/diffusion.yaml

@@ -5,12 +5,12 @@ data:
   sampling_rate: 44100
   block_size: 512 # Equal to hop_length
   duration: 2 # Audio duration during training, must be less than the duration of the shortest audio clip
-  encoder: 'hubertsoft' # 'hubertsoft', 'hubertbase', 'hubertbase768', 'contentvec', 'contentvec768' or 'contentvec768l12' or 'cnhubertsoftfish'
+  encoder: 'contentvec' # 'hubertsoft', 'hubertbase', 'hubertbase768', 'contentvec', 'contentvec768' or 'contentvec768l12' or 'cnhubertsoftfish'
   cnhubertsoft_gate: 10 # only use with cnhubertsoftfish
   encoder_sample_rate: 16000
   encoder_hop_size: 320
   encoder_out_channels: 256 # 768 if using 'hubertbase768', 'contentvec768' or 'contentvec768l12' encoder
-  encoder_ckpt: pretrain/hubert/hubert-soft-0d54a1f4.pt
+  encoder_ckpt: /workspace/DDSP-SVC/pretrain/hubert/checkpoint_best_legacy_500.pt
   train_path: data/train # Create a folder named "audio" under this path and put the audio clip in it
   valid_path: data/val # Create a folder named "audio" under this path and put the audio clip in it
 model:
@@ -23,7 +23,7 @@ model:
 device: cuda
 vocoder:
   type: 'nsf-hifigan'
-  ckpt: 'pretrain/nsf_hifigan/model'
+  ckpt: '/workspace/DDSP-SVC/pretrain/nsf_hifigan/model'
 infer:
   speedup: 10
   method: 'dpm-solver' # 'pndm' or 'dpm-solver'
@@ -31,16 +31,16 @@ env:
   expdir: exp/diffusion-test
   gpu_id: 0
 train:
-  num_workers: 2 # If your cpu and gpu are both very strong, set to 0 may be faster!
-  amp_dtype: fp32 # fp32, fp16 or bf16 (fp16 or bf16 may be faster if it is supported by your gpu)
-  batch_size: 48
+  num_workers: 0 # If your cpu and gpu are both very strong, set to 0 may be faster!
+  amp_dtype: fp16 # fp32, fp16 or bf16 (fp16 or bf16 may be faster if it is supported by your gpu)
+  batch_size: 96
   cache_all_data: true # Save Internal-Memory or Graphics-Memory if it is false, but may be slow
-  cache_device: 'cpu' # Set to 'cuda' to cache the data into the Graphics-Memory, fastest speed for strong gpu
+  cache_device: 'cuda' # Set to 'cuda' to cache the data into the Graphics-Memory, fastest speed for strong gpu
   cache_fp16: true
   epochs: 100000
   interval_log: 10
   interval_val: 2000
-  interval_force_save: 20000
+  interval_force_save: 10000
   lr: 0.0002
   decay_step: 100000
   gamma: 0.5