search

uber-archive / plato-research-dialogue-system

This is the Plato Research Dialogue System, a flexible platform for developing conversational AI agents.
Apache License 2.0
979 stars 191 forks source link

How to Train with Deep Learning Framework? #3

Closed ayaz-amin closed 4 years ago

ayaz-amin commented 4 years ago

I read the docs and noticed that models are currently trainable using Ludwig (although it mentioned that models can be trained in Tensorflow or PyTorch as well). I want to be able to train models using a DL framework. Is it doable right now?

chagri commented 4 years ago

Plato is compatible with any Python Library. You can extend Plato with any library of your choice: TF, PyTorch, Keras, etc. by writing custom code.

On Sun, Aug 4, 2019 at 4:14 PM Ayaz Amin notifications@github.com wrote:

I read the docs and noticed that models are currently trainable using Ludwig. I want to be able to train models using a DL framework (i.e. Tensorflow and PyTorch). Is it doable right now?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_uber-2Dresearch_plato-2Dresearch-2Ddialogue-2Dsystem_issues_3-3Femail-5Fsource-3Dnotifications-26email-5Ftoken-3DABMEMFMSWHGY5RV5BSOTMB3QC5PG5A5CNFSM4IJGOQY2YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HDINDDA&d=DwMCaQ&c=r2dcLCtU9q6n0vrtnDw9vg&r=5xfItXPi0scrEEcywmVDYv0OcHvfFjioaELjPY4TT1g&m=RpUB-pHUAKc7f8eUAPN3AfB2OGdx7bmjzEp8ZC0x3HQ&s=75qGu7_9o1Uvz0lrLhoP32HvPi5fWHJSY3GADBVjEYU&e=, or mute the thread https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_notifications_unsubscribe-2Dauth_ABMEMFIEJBOZ43RCQ2IR2S3QC5PG5ANCNFSM4IJGOQYQ&d=DwMCaQ&c=r2dcLCtU9q6n0vrtnDw9vg&r=5xfItXPi0scrEEcywmVDYv0OcHvfFjioaELjPY4TT1g&m=RpUB-pHUAKc7f8eUAPN3AfB2OGdx7bmjzEp8ZC0x3HQ&s=9qp3A6advsPWlCKdGiOgb1qkD-6sBRSd0i-3AsRbZQM&e= .

ayaz-amin commented 4 years ago

Thanks for replying! I will try to use Plato with Keras, and if successful, create an extension for Plato.

kouohhashi commented 4 years ago

@ayaz-amin Hi, have you use Tensorflow or Pytorch with Plato? If so could you share tips here or blog?

I think tutorial with TF or Pytorch is convenient for many developers because most developers use TF or Pytorch. Ludwig is not for developers...

w4nderlust commented 4 years ago

@kouohhashi I suggest you to take a closer look at Ludwig. It covers all dialogue systems component modeling use cases, and if you can be extremely easily extend it for any use case you have in mind as it provides a tensor-in tensor-out interface inside which you can put whatever model you want, even though the most common ones are already available to you as they have already been implemented.