#Not working with MacOs Sonoma
63 messages · Page 1 of 1 (latest)
For file inference or realtime?
Both, it's primarily an issue with the mps backend
it's already solved
exactly
how to solve?
I also meet the same question
the voice sound terrible with MacOs Sonoma
@elfin stratus
inferences or training
I can hear it
is imput
2 months ago is ok
It sounds good when I used RVC 2 months ago
I think it's in imput
Running on local URL: http://0.0.0.0:7865
loading weights/Elvis50v2.pth
gin_channels: 256 self.spk_embed_dim: 109
<All keys matched successfully>
2023-11-04 15:08:13 | INFO | fairseq.tasks.hubert_pretraining | current directory is /Users/wendao/RVC
2023-11-04 15:08:13 | INFO | fairseq.tasks.hubert_pretraining | HubertPretrainingTask Config {'_name': 'hubert_pretraining', 'data': 'metadata', 'fine_tuning': False, 'labels': ['km'], 'label_dir': 'label', 'label_rate': 50.0, 'sample_rate': 16000, 'normalize': False, 'enable_padding': False, 'max_keep_size': None, 'max_sample_size': 250000, 'min_sample_size': 32000, 'single_target': False, 'random_crop': True, 'pad_audio': False}
2023-11-04 15:08:13 | INFO | fairseq.models.hubert.hubert | HubertModel Config: {'_name': 'hubert', 'label_rate': 50.0, 'extractor_mode': default, 'encoder_layers': 12, 'encoder_embed_dim': 768, 'encoder_ffn_embed_dim': 3072, 'encoder_attention_heads': 12, 'activation_fn': gelu, 'layer_type': transformer, 'dropout': 0.1, 'attention_dropout': 0.1, 'activation_dropout': 0.0, 'encoder_layerdrop': 0.05, 'dropout_input': 0.1, 'dropout_features': 0.1, 'final_dim': 256, 'untie_final_proj': True, 'layer_norm_first': False, 'conv_feature_layers': '[(512,10,5)] + [(512,3,2)] * 4 + [(512,2,2)] * 2', 'conv_bias': False, 'logit_temp': 0.1, 'target_glu': False, 'feature_grad_mult': 0.1, 'mask_length': 10, 'mask_prob': 0.8, 'mask_selection': static, 'mask_other': 0.0, 'no_mask_overlap': False, 'mask_min_space': 1, 'mask_channel_length': 10, 'mask_channel_prob': 0.0, 'mask_channel_selection': static, 'mask_channel_other': 0.0, 'no_mask_channel_overlap': False, 'mask_channel_min_space': 1, 'conv_pos': 128, 'conv_pos_groups': 16, 'latent_temp': [2.0, 0.5, 0.999995], 'skip_masked': False, 'skip_nomask': False, 'checkpoint_activations': False, 'required_seq_len_multiple': 2, 'depthwise_conv_kernel_size': 31, 'attn_type': '', 'pos_enc_type': 'abs', 'fp16': False}
@elfin stratus
it doesn't give any error
change it to mangio or others
use it in mangio crepe and see how it looks in audio
what's the mangio crepe?
send a photo of the interface
it still sound terrible
I changed to Mangio Crepé RVC2
It's the Macos Sonoma's issue
let me read that
pip3 --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/cpu
I have the same issue. Someone with MacOs Sonoma that have a solution for this?
pip3 --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/cpu
put this on your terminal
hit enter
I already done that on sonoma 14.0 but it did not worked
It will work on Sonoma 14.1?
@elfin stratus the command you mention is this one right?
pip3 install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/cpu
pip3 install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/cpu
yes
tested also with the Sonoma. I will try the Mangio Crepé RVC2
try it and see how the inference sounds
didn't works. Just a terrible output also with the RVC2
@old ocean did you figure out how to solve this?
still the same
well after that it started working with old trained models generated before the upgrade to Sonoma. With new trained models the inferences are just moaning sounds