search

sp-uhh / sgmse

Score-based Generative Models (Diffusion Models) for Speech Enhancement and Dereverberation
MIT License
454 stars 69 forks source link

The enhanced.py can't generate the enhanced.wav #45

Closed Lieber0402 closed 2 months ago

Lieber0402 commented 5 months ago

100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 3/3 [26:19<00:00, 526.42s/it]

While this is showing in the terminal, nothing genertated in the target directory.

This is the command to excute the enhanced.py: python enhancement.py --test_dir D:\Desktop\11\train --enhanced_dir D:\Desktop\processed\denoise --ckpt D:\Desktop\pre_trained\denoise\train_vb_29nqe0uh_epoch=115.ckpt

Nothing was generated in the --enhanced_dir.

PavelPanjaya commented 4 months ago

Same here.

Their code assumes there is a noisy directory inside your test directory:

noisy_dir = join(args.test_dir, 'noisy/')
julius-richter commented 2 months ago

Hi there,

I have updated enhancement.py. Now --test_dir should point to the directory containing the test data (without assuming a noisy directory inside). --enhanced_dir is the directory where the enhanced files will be saved.