Path to trained files, checkpoint, etc passed by an argument about nmt-chatbot

[daniel-kukiela]

I think that's a good idea.
As a workaround you can edit model/checkpoint file.
I'm not sure how did you get filename like that.

[binksdata]

Yes, editing the checkpoint file works. That file name is like names from Sentdex tutorial, look quite similar.

https://pythonprogramming.net/interacting-chatbot-deep-learning-python-tensorflow/

"checkpoint
translate.ckpt-225000.meta
translate.ckpt-225000.index
translate.ckpt-225000.data-00000-of-00001
"

[daniel-kukiela]

I've added ability to pass checkpoint name as an argument, for example:

python inference.py translate.ckpt-1000

I'll add deployment script as well soon (it's almost redy) prepairing model for deployment including ability to choose model (it will save all necessary files to run inference in separate folder).

[daniel-kukiela]

Deployment utility added to repository.