After a fresh install of torch7 and the unsup rocks, running the train-autoencoder (th train-autoencoder.lua) demonstration ends with the following error
==> constructed convolutional predictive sparse decomposition (PSD) auto-encoder
/usr/local/bin/luajit: /usr/local/share/lua/5.1/unsup/psd.lua:39: attempt to call method 'initDiagHessianParameters' (a nil value)
stack traceback:
/usr/local/share/lua/5.1/unsup/psd.lua:39: in function 'initDiagHessianParameters'
train-autoencoder.lua:225: in main chunk
[C]: in function 'dofile'
/usr/local/lib/luarocks/rocks/trepl/scm-1/bin/th:108: in main chunk
[C]: at 0x00404ec4
After a fresh install of torch7 and the unsup rocks, running the train-autoencoder (
th train-autoencoder.lua
) demonstration ends with the following errorany idea on how to fix this ?