a706892b59
- remove support for cuDNN - add new pre-trained model and json files - some changes in training script If you have cuDNN model, Please run following commands to convert. $ cp models/your_own_model.t7 models/your_own_model.t7.backup $ th cudnn2cunn.lua -model models/your_own_model.t7
11 lines
568 B
Bash
Executable file
11 lines
568 B
Bash
Executable file
#!/bin/sh
|
|
|
|
th train.lua -method noise -noise_level 1 -model_dir models/anime_style_art -test images/miku_noisy.png
|
|
th cleanup_model.lua -model models/anime_style_art/noise1_model.t7 -oformat ascii
|
|
|
|
th train.lua -method noise -noise_level 2 -model_dir models/anime_style_art -test images/miku_noisy.png
|
|
th cleanup_model.lua -model models/anime_style_art/noise2_model.t7 -oformat ascii
|
|
|
|
th train.lua -method scale -scale 2 -model_dir models/anime_style_art -test images/miku_small.png
|
|
th cleanup_model.lua -model models/anime_style_art/scale2.0x_model.t7 -oformat ascii
|