23 lines
2.3 KiB
Bash
Executable file
23 lines
2.3 KiB
Bash
Executable file
#!/bin/sh
|
|
|
|
# convert data
|
|
th convert_data.lua -max_training_image_size 1600
|
|
|
|
# scale
|
|
# after train, I selected `scale2.0x_model.86-1.t7`, because seemingly this model is best mesh-artifact free model. Maybe visual testing is required.
|
|
# ref: https://github.com/nagadomi/waifu2x/issues/125
|
|
th train.lua -save_history 1 -scale 2 -model upconv_7 -method scale -model_dir models/test/upconv_7_rev5 -downsampling_filters "Box,Sinc" -test query/pixel-art-small.png -backend cudnn -thread 4 -oracle_rate 0.0 -inner_epoch 2 -epoch 100
|
|
|
|
# noise_scale
|
|
|
|
th train.lua -save_history 1 -model upconv_7 -method noise_scale -noise_level 0 -model_dir models/test/upconv_7_rev6 -downsampling_filters "Box,Sinc" -test query/noise_test.jpg -backend cudnn -thread 4 -resume models/test/upconv_7_rev5/scale2.0x_model.t7 -style art
|
|
th train.lua -save_history 1 -model upconv_7 -method noise_scale -noise_level 1 -model_dir models/test/upconv_7_rev6 -downsampling_filters "Box,Sinc" -test query/noise_test.jpg -backend cudnn -thread 4 -resume models/test/upconv_7_rev5/scale2.0x_model.t7 -style art
|
|
th train.lua -save_history 1 -model upconv_7 -method noise_scale -noise_level 2 -model_dir models/test/upconv_7_rev6 -downsampling_filters "Box,Sinc" -test query/noise_test.jpg -backend cudnn -thread 4 -resume models/test/upconv_7_rev5/scale2.0x_model.t7 -style art
|
|
th train.lua -save_history 1 -model upconv_7 -method noise_scale -noise_level 3 -model_dir models/test/upconv_7_rev6 -downsampling_filters "Box,Sinc" -test query/noise_test.jpg -backend cudnn -thread 4 -resume models/test/upconv_7_rev5/scale2.0x_model.t7 -style art -nr_rate 1
|
|
|
|
# noise
|
|
th train.lua -save_history 1 -model vgg_7 -method noise -noise_level 0 -model_dir models/test/yuv420_rev2 -test query/noise_test.jpg -backend cudnn -thread 4 -style art -crop_size 32
|
|
th train.lua -save_history 1 -model vgg_7 -method noise -noise_level 1 -model_dir models/test/yuv420_rev2 -test query/noise_test.jpg -backend cudnn -thread 3 -style art -crop_size 32
|
|
th train.lua -save_history 1 -model vgg_7 -method noise -noise_level 2 -model_dir models/test/yuv420_rev2 -test query/noise_test.jpg -backend cudnn -thread 3 -style art -crop_size 32
|
|
th train.lua -save_history 1 -model vgg_7 -method noise -noise_level 3 -model_dir models/test/yuv420_rev2 -test query/noise_test.jpg -backend cudnn -thread 3 -style art -crop_size 32 -nr_rate 1
|