run_segan+_clean.sh

#!/bin/bash

CKPT_PATH="ckpt_segan+"

# please specify the path to your G model checkpoint
# as in weights_G-EOE_<iter>.ckpt
G_PRETRAINED_CKPT="segan+_generator.ckpt"

# please specify the path to your folder containing
# noisy test files, each wav in there will be processed
TEST_FILES_PATH="data_veu4/expanded_segan1_additive/noisy_testset/"

# please specify the output folder where cleaned files
# will be saved
SAVE_PATH="synth_segan+"

python -u clean.py --g_pretrained_ckpt $CKPT_PATH/$G_PRETRAINED_CKPT \
	--test_files $TEST_FILES_PATH --cfg_file $CKPT_PATH/train.opts \
	--synthesis_path $SAVE_PATH --soundfile