command.txt

--file-list ./speaker1/speaker1-train-files.txt --in-dir ./speaker1/p225 --out ./speaker1/vctk-speaker1-train-long.h5 --scale_low 16 --scale_high 4 --sr 16000 --dimension 32000

--file-list ./speaker1/speaker1-val-files.txt --in-dir ./speaker1/p225 --out ./speaker1/vctk-speaker1-val-long.h5 --scale_low 16 --scale_high 4 --sr 16000 --dimension 32000

train --train ../data/vctk/speaker1/vctk-speaker1-train-long.h5   --val ../data/vctk/speaker1/vctk-speaker1-val-long.h5   -e 120   --batch-size 64   --lr 3e-4   --logname singlespeaker

eval --logname singlespeaker.lr0.000020.1.g4.b64.2/model.ckpt-40001 --out-label singlespeaker-out --wav-file-list ../data/vctk/speaker1/speaker1-val-files.txt --r 4

0225:

--file-list ./speaker1/speaker1-train-files.txt --in-dir ./speaker1/p225 --out ./speaker1/vctk-speaker1-train.h5 --scale_low 16 --scale_high 4 --sr 16000

--file-list ./speaker1/speaker1-val-files.txt --in-dir ./speaker1/p225 --out ./speaker1/vctk-speaker1-val.h5 --scale_low 16 --scale_high 4 --sr 16000

train --train ../data/vctk/speaker1/vctk-speaker1-train.h5   --val ../data/vctk/speaker1/vctk-speaker1-val.h5   -e 800   --batch-size 64   --lr 1e-4   --logname original_model_single_speaker

eval --logname original_model_4_1k_4k/model.ckpt-52801 --out-label singlespeaker-standard --wav-file-list ../data/vctk/speaker1/speaker1-val-files.txt --r 4

0306
CUDA_VISIBLE_DEVICES=0 python run.py train --train vctk-multispeaker-train.4.16.16000.32768.8192.1.h5   --val vctk-multispeaker-val.4.16.16000.32768.8192.1.h5   -e 240   --batch-size 64   --lr 3e-4   --logname original_model_full_vctk

python run.py train --train vctk-multispeaker-train.4.16.16000.32768.8192.1.h5   --val vctk-multispeaker-val.4.16.16000.32768.8192.1.h5   -e 240   --batch-size 128   --lr 3e-4   --logname original_model_full_vctk_continued  --from_ckpt model.ckpt-25103

eval --logname model.ckpt-25103 --out-label singlespeaker-standard --wav-file-list ../data/vctk/speaker1/speaker1-val-files.txt --r 4

0325
python run.py eval --logname model.ckpt-25103 --out-label multispeaker-original-conti --wav-file-list ../data/vctk/speaker1/speaker1-val-files.txt --r 4