Error running neuraltalk2 on Ubuntu

Asked by Parag Ahivale on 2015-12-20

I get the following error while executing the following

parag@parag:~/MyHome/neuraltalk2/neuraltalk2-master$ th eval.lua -model /home/parag/MyHome/neuraltalk2/ -image_folder /home/parag/MyHome/neuraltalk2/images/ -num_images -1 -gpuid -1
/home/parag/torch/install/bin/luajit: /home/parag/torch/install/share/lua/5.1/torch/File.lua:199: read error: read 0 blocks instead of 1 at /home/parag/torch/pkg/torch/lib/TH/THDiskFile.c:314
stack traceback:
 [C]: in function 'readInt'
 /home/parag/torch/install/share/lua/5.1/torch/File.lua:199: in function 'readObject'
 /home/parag/torch/install/share/lua/5.1/torch/File.lua:325: in function 'load'
 eval.lua:69: in main chunk
 [C]: in function 'dofile'
 ...arag/torch/install/lib/luarocks/rocks/trepl/scm-1/bin/th:145: in main chunk
 [C]: at 0x00405d70
parag@parag:~/MyHome/neuraltalk2/neuraltalk2-master$

Can someone help here?

Question information

Language:
English Edit question
Status:
Solved
For:
Ubuntu totem Edit question
Assignee:
No assignee Edit question
Solved by:
Parag Ahivale
Solved:
2015-12-21
Last query:
2015-12-21
Last reply:
2015-12-20

How did you install the application please?

Parag Ahivale (ahivale-parag) said : #2

$ curl -s https://raw.githubusercontent.com/torch/ezinstall/master/install-deps | bash
$ git clone https://github.com/torch/distro.git ~/torch --recursive
$ cd ~/torch;
$ ./install.sh # and enter "yes" at the end to modify your bashrc
$ source ~/.bashrc
$ luarocks install nn
$ luarocks install nngraph
$ luarocks install image
sudo apt-get install libhdf5-serial-dev hdf5-tools
git clone <email address hidden>:deepmind/torch-hdf5.git
cd torch-hdf5
luarocks make hdf5-0-0.rockspec LIBHDF5_LIBDIR="/usr/lib/x86_64-linux-gnu/"
luarocks install loadcaffe

Parag Ahivale (ahivale-parag) said : #3

This command solved the issue for me:

$ th eval.lua -model /home/parag/MyHome/neuraltalk2/model_id1-501-1448236541.t7_cpu.t7 -image_folder /home/parag/MyHome/neuraltalk2/images/ -num_images -1 -gpuid -1

All I have done here is included the model name in the path instead of writing just the path.