fast-neural-style
fast-neural-style copied to clipboard
hdf5/group.lua:88: attempt to concatenate 'int64_t' and 'string'hdf5/
Hi, after training a model by
python scripts/make_style_dataset.py \```
--train_dir ~/packs-fns/train \
--val_dir ~/packs-fns/valid \
--output_file dataset1.h5
I train the model by
th train.lua \
-h5_file dataset1.h5 \
-style_image ~/packs-fns/train/wisegeek.jpg \
-style_image_size 384 \
-content_weights 1.0 \
-style_weights 5.0 \
-checkpoint_name checkpoint \
-gpu 0
This throws the following error message I am unable to understand:
/home/rstudio/torch/install/bin/luajit: /home/rstudio/torch/install/share/lua/5.1/hdf5/group.lua:88: attempt to concatenate 'int64_t' and 'string'
stack traceback:
/home/rstudio/torch/install/share/lua/5.1/hdf5/group.lua:88: in function 'tostring'
/home/rstudio/torch/install/share/lua/5.1/hdf5/group.lua:39: in function '__init'
/home/rstudio/torch/install/share/lua/5.1/torch/init.lua:91: in function </home/rstudio/torch/install/share/lua/5.1/torch/init.lua:87>
[C]: in function 'HDF5Group'
/home/rstudio/torch/install/share/lua/5.1/hdf5/init.lua:74: in function '_loadObject'
/home/rstudio/torch/install/share/lua/5.1/hdf5/file.lua:19: in function '__init'
/home/rstudio/torch/install/share/lua/5.1/torch/init.lua:91: in function </home/rstudio/torch/install/share/lua/5.1/torch/init.lua:87>
[C]: in function 'open'
./fast_neural_style/DataLoader.lua:17: in function '__init'
/home/rstudio/torch/install/share/lua/5.1/torch/init.lua:91: in function </home/rstudio/torch/install/share/lua/5.1/torch/init.lua:87>
[C]: in function 'DataLoader'
train.lua:138: in function 'main'
I couldn't find any hint by googling so I'm getting desperate. Any ideas or hints??
Apply the following change to the hdf5 module: https://github.com/adynathos/torch-hdf5/commit/539fc15ecd25b93f4d0784ea4eb75681329f9714