Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

the missing full source code of plato-2 #32

Closed
chikiuso opened this issue Nov 20, 2020 · 9 comments
Closed

the missing full source code of plato-2 #32

chikiuso opened this issue Nov 20, 2020 · 9 comments

Comments

@chikiuso
Copy link

Hi thanks for your great work! I explore the plato-2 directory and just found there are .sh files, may I ask where is the .py files? so I could try the chatbot interaction, thanks for your help!

@sserdoubleh
Copy link
Collaborator

You can run git clone https://github.com/PaddlePaddle/Knover.git
PLATO-2 model implement is in models/plato.py
If you want to run interaction, you can run scripts fllowing plato-2/README.md

@chikiuso
Copy link
Author

chikiuso commented Nov 20, 2020

@sserdoubleh thanks for your help, I run it finally, but it halts on this screen :

https://ibb.co/VpbQwp2

@sserdoubleh
Copy link
Collaborator

@sserdoubleh thanks for your help, I run it finally, but it halts on this screen :

https://ibb.co/VpbQwp2

What about the version of PaddlePaddle?

@sserdoubleh
Copy link
Collaborator

#11
Perhaps because of the version of CUDA, PaddlePaddle cannot support CUDA 11.0 now.

@chikiuso
Copy link
Author

@sserdoubleh thanks for your help, I work now. is this project possible to add persona or personality to the chat bot? if yes, what procedure should I follow to fine-tune it to a personalized chat bot? thanks.

@sserdoubleh
Copy link
Collaborator

@sserdoubleh thanks for your help, I work now. is this project possible to add persona or personality to the chat bot? if yes, what procedure should I follow to fine-tune it to a personalized chat bot? thanks.

#24 (comment)
This may be helpful.

If you want to fine-tune your model with persona, you can check the data formt in data/train.tsv and the description in README.md on the root path.

@chikiuso
Copy link
Author

Hi @sserdoubleh, thanks for your help, I now could run the 24L model interact on server, is it possible to make it work with cuda 11.0 or above? or could you point some way I could try to modify the code to make it cuda 11.0 compatible? as the latest gpu all run cuda 11.0 or above.

thanks for your great help in the last several days, your responsive reply make it a successful paper and project.

@sserdoubleh
Copy link
Collaborator

sserdoubleh commented Nov 24, 2020

@chikiuso I have sent the feedback to PaddlePaddle. Maybe it need time to support CUDA11.0. If it has any progress, I will notice you.

May be you can try to re-compile your PaddlePadle~
https://www.paddlepaddle.org.cn/documentation/docs/en/install/compile/compile_Ubuntu_en.html
git checkout release/1.8

@chikiuso
Copy link
Author

Hi @sserdoubleh , thanks for your help, I just compiled paddlepaddle from source and have a try but failed , may be I wait for the paddlepaddle upgrade support, thanks a lot!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants