Cannot load released model

Support
1

th translate.lua -model model_epoch13_1226.32_release.t7 -src data/src-test.txt -output pred.txt

I get the following error when i try and run the above command. Please help.

[06/27/17 13:30:07 INFO] Loading 'model_epoch13_1226.32_release.t7'...	
/Users/shivaligoel/torch/install/bin/luajit: ...s/shivaligoel/torch/install/share/lua/5.1/torch/File.lua:375: unknown object
stack traceback:
	[C]: in function 'error'
	...s/shivaligoel/torch/install/share/lua/5.1/torch/File.lua:375: in function 'readObject'
	...s/shivaligoel/torch/install/share/lua/5.1/torch/File.lua:368: in function 'readObject'
	...rs/shivaligoel/torch/install/share/lua/5.1/nn/Module.lua:192: in function 'read'
	...s/shivaligoel/torch/install/share/lua/5.1/torch/File.lua:351: in function 'readObject'
	...s/shivaligoel/torch/install/share/lua/5.1/torch/File.lua:369: in function 'readObject'
	...rs/shivaligoel/torch/install/share/lua/5.1/nn/Module.lua:192: in function 'read'
	...s/shivaligoel/torch/install/share/lua/5.1/torch/File.lua:351: in function 'readObject'
	...s/shivaligoel/torch/install/share/lua/5.1/torch/File.lua:369: in function 'readObject'
	...s/shivaligoel/torch/install/share/lua/5.1/torch/File.lua:369: in function 'readObject'
	...s/shivaligoel/torch/install/share/lua/5.1/torch/File.lua:369: in function 'readObject'
	...
	...s/shivaligoel/torch/install/share/lua/5.1/torch/File.lua:369: in function 'readObject'
	...s/shivaligoel/torch/install/share/lua/5.1/torch/File.lua:369: in function 'readObject'
	...s/shivaligoel/torch/install/share/lua/5.1/torch/File.lua:409: in function 'load'
	./onmt/translate/Translator.lua:117: in function '__init'
	...s/shivaligoel/torch/install/share/lua/5.1/torch/init.lua:91: in function 'new'
	translate.lua:53: in function 'main'
	translate.lua:196: in main chunk
	[C]: in function 'dofile'
	...goel/torch/install/lib/luarocks/rocks/trepl/scm-1/bin/th:150: in main chunk
	[C]: at 0x010bbb09f0
2

Is it a 32 bits system by any chance?

If not, could you also share the options you used for training?

3

No, its not. I’m running it on my mac.

For training I used the default parameters on a GPU.

4

I can only think that the file is corrupted somehow. Try running the release script or download the model again.

5

@Shivali, did you solve this issue?
I had the same problem trying to run a released model trained using GPU.

Thanks!

6

@welingtonveiga Yes the issue was resolved. I wasn’t able to run it on my Macbook but the same ran well on a Linux Machine. I’m still wondering why that happened, because I can run CPU trained models on my laptop but not this released one.

7

Thanks @Shivali, in my case the two machine uses Linux (distinct distributions only) and after your feedback I tried to test on the same machine, but this issue persists.