OutOfMemory while loading a GPU model?

Support
1

I’m trying to deploy on a new machine. While loading a GPU model, I get this error:

[06/02/1    7 12:00:31 INFO] Loading '/home/onmt/MODELS/MANAGEMENT/onmt-fr-en-modelSTDG_epoch35.t7'...    
    /home/onmt/torch/install/bin/luajit: /home/onmt/torch/install/share/lua/5.1/torch/File.lua:272: $ Torch: invalid memory size -- maybe an overflow? at /home/onmt/torch/pkg/torch/lib/TH/THGeneral.c:257
    stack traceback:
        [C]: in function 'readChar'
        /home/onmt/torch/install/share/lua/5.1/torch/File.lua:272: in function 'readObject'
        /home/onmt/torch/install/share/lua/5.1/torch/File.lua:368: in function 'readObject'
        /home/onmt/torch/install/share/lua/5.1/torch/File.lua:369: in function 'readObject'
        /home/onmt/torch/install/share/lua/5.1/torch/File.lua:353: in function 'readObject'
        /home/onmt/torch/install/share/lua/5.1/torch/File.lua:369: in function 'readObject'
        /home/onmt/torch/install/share/lua/5.1/torch/File.lua:369: in function 'readObject'
        /home/onmt/torch/install/share/lua/5.1/torch/File.lua:369: in function 'readObject'
        /home/onmt/torch/install/share/lua/5.1/torch/File.lua:409: in function 'load'
        ./onmt/translate/Translator.lua:34: in function '__init'
        /home/onmt/torch/install/share/lua/5.1/torch/init.lua:91: in function 'new'
        tools/translation_server.lua:77: in function 'main'
        tools/translation_server.lua:109: in main chunk
        [C]: in function 'dofile'
        ...onmt/torch/install/lib/luarocks/rocks/trepl/scm-1/bin/th:150: in main chunk
        [C]: at 0x00405d50

Looking at File.lua:272, I suspect the read size to be wrong…(!?)

:neutral_face:

2

Looks like it read a negative size in the model. Is your file corrupted in some way?

1 Like
3

Yes !! Retrieving the model in a new fresh copy did it !

Thanks !
:smiley: