Open gigix opened 7 years ago
Hi there,
I wonder if it possible to use a trained model in another way - rather than generating new sample text, I'd like to input some new text and ask the model if the new text looks similar to the training input. The interface might look like this...
$ th grade.lua -checkpoint cv/some_checkpoint.t7 -input new_text.txt Similarity: 85%
$ th grade.lua -checkpoint cv/some_checkpoint.t7 -input new_text.txt
Similarity: 85%
Hi there,
I wonder if it possible to use a trained model in another way - rather than generating new sample text, I'd like to input some new text and ask the model if the new text looks similar to the training input. The interface might look like this...
$ th grade.lua -checkpoint cv/some_checkpoint.t7 -input new_text.txt
Similarity: 85%