Atma
> Could you try to create a new conda environment and install ConvLab-2 again? We don't have a constraint on the version of boto3 & botocore. In the [Colab](https://colab.research.google.com/github/thu-coai/ConvLab-2/blob/master/tutorials/Getting_Started.ipynb), I...
Tried with python 3.7.11 and got new error: ``` Failed to build jsonnet quadprog ERROR: Could not build wheels for quadprog which use PEP 517 and cannot be installed directly...
> I succeed to fix all error, by (1) Use a higher version of python. (2) Fix the installation of jsonnet quadprog by directly installing them with conda.
> maybe update the pip will solve the problem, too emmmm.. The problem seems coming from the deep dependency crash caused by python version, since I try many different version...
You might be right, but the teacher forcing here could really improve the performance of 1 ~ 2 point~
Yep~~ You could have a try to tune the teacher forcing rate (default 0), 0.5 is worth trying. I found both 0 and 1 are not helping. emmmmm..... From my...
@umgupta Ha~Your post has also deepened my understanding of teacher forcing. Maybe I should implement a kind of teacher forcing that you pointed out, which can further enhance my model...
@umgupta Machine translation problem in pytorch tutorial is quite simple, which might satisfy you.
Thanks for your reply, it will be in great need.
这个版本的code没有测试过多GPU，目前只能是先用单GPU跑 ：（