search

santi-pdp / segan

Speech Enhancement Generative Adversarial Network in TensorFlow
MIT License
816 stars 281 forks source link

Could it be in real time ? #69

Open ucasiggcas opened 5 years ago

ucasiggcas commented 5 years ago

Hi, If I use the trained model, will I get the denoised wav in real time ?

Thx

ucasiggcas commented 5 years ago

When I test with CUDA_VISIBLE_DEVICES="" python main.py --init_noise_std 0. --save_path ckpt --batch_size 100 --g_nl prelu --weights SEGAN_full --test_wav mic_01.wav --clean_save_path ./ I have the same problem,who can help me ? https://github.com/santi-pdp/segan/issues/68

ucasiggcas commented 5 years ago

ValueError: Dimension 0 in both shapes must be equal, but are 1 and 16383. Shapes are [1] and [16383]. From merging shape 0 with other shapes. for 'device_0/concat/concat_dim' (op: 'Pack') with input shapes: [1], [16383].

fighterG commented 5 years ago

I also met this problem. Have you solved it

ucasiggcas commented 5 years ago

I also met this problem. Have you solved it

pls see here https://github.com/santi-pdp/segan/issues/68

fighterG commented 5 years ago

I also met this problem. Have you solved it

pls see here

68

thank you

yannier912 commented 3 years ago

@ucasiggcas hello, did you solve the problem about real-time? it takes too long time...