PS G:\DDSP-SVC> python train.py -c configs/combsub.yaml
> config: configs/combsub.yaml
> exp: exp/combsub-test
[DDSP Model] Combtooth Subtractive Synthesiser
[*] restoring model from exp/combsub-test\model_300000.pt
Traceback (most recent call last):
File "train.py", line 68, in <module>
initial_global_step, model, optimizer = utils.load_model(args.env.expdir, model, optimizer, device=args.device)
File "G:\DDSP-SVC\logger\utils.py", line 119, in load_model
model.load_state_dict(ckpt['model'])
File "C:\Users\29099\.virtualenvs\DDSP-SVC-YOgpXN-h\lib\site-packages\torch\nn\modules\module.py", line 2041, in load_state_dict
raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format(
RuntimeError: Error(s) in loading state_dict for CombSubFast:
Unexpected key(s) in state_dict: "unit2ctrl.spk_embed.weight".
配置文件:
data:
f0_extractor: 'parselmouth' # 'parselmouth', 'dio', 'harvest', or 'crepe'
f0_min: 65 # about C2
f0_max: 800 # about G5
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' or 'contentvec'
encoder_sample_rate: 16000
encoder_hop_size: 320
encoder_out_channels: 256
encoder_ckpt: pretrain/hubert/hubert-soft-0d54a1f4.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:
type: 'CombSubFast'
n_spk: 1 # max number of different speakers
enhancer:
type: 'nsf-hifigan'
ckpt: 'pretrain/nsf_hifigan/model'
loss:
fft_min: 256
fft_max: 2048
n_scale: 4 # rss kernel numbers
device: cuda
env:
expdir: exp/combsub-test
gpu_id: 0
train:
num_workers: 0 # If your cpu and gpu are both very strong, set to 0 may be faster!
batch_size: 24
cache_all_data: true # Save Internal-Memory or Graphics-Memory if it is false, but may be slow
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
lr: 0.0005
weight_decay: 0
配置文件: