Philip May
Philip May
What do others think about this? @fchollet @karmel , ...
Hello dear maintainers of this project. What about this idea? What do you think? Thanks Philip
> you can use **_strict=False_** in load_state_dict. This can solved the issue. > > ``` > model.load_state_dict(checkpoint['state_dict'], strict=False) > ``` Well - does it solve the issue or just hide...
Hey @stefan-it - good luck with getting an answer. I think the google research guys have pretty much abandoned this repo. @Phil1108 I think we never had a problem like...
Or this: https://github.com/richarddwang/electra_pytorch
> I used a sentence-segmented training corpus. That means, each line of the input file contains one sentence :) Is that one sentence a line thing ok? I am asking...
Can this be closed since the PR was merged?
Good job @blakechi ! Many thanks.
Well, I made a branch where I can enable normalize. I did some tests with default settings, a normal BERT model (no pretrained sentence embedding model) and optuna. Letting optuna...
@lewtun what is your opinion on this? Do you have more insights?