Tae Hwan Jung

> Regarding the solution, you do not need to do anything :-). We need to implement the solution I suggested inside [TileDB core](https://github.com/TileDB-Inc/TileDB) and once we merge a patch, you...

@lukemelas Hello Thank you for your interest in my project. I hope finetuning with PyTorch but it can't not now because we convert tensorflow checkpoint(original github XLNet) to pytorch model...

There are no code releated to specific tasks about inference or fine-tuing :( if you want to see more recommend here : https://github.com/huggingface/transformers Thanks

Pre-training is Yes. But fine-tuning task is not Yet

@yuanjungod Could you tell me where is wrong point?? @xingchensong Thanks for your interest in my project. !! No not yet huggingface already did it! https://github.com/huggingface/transformers < in here You...

I think it is not important accuracy in this domain

No it was not implemented :( :( Please contribute for that!

Hello I am also trying training the torchmoji without pre-trained model. but I have problem that loss is not decrease well....

Hello mhasan1, I solved this problem From ` best_loss = np.mean([calc_loss(loss_op, model(Variable(xv)), Variable(yv)).data.cpu().numpy()[0] for xv, yv in val_gen])` To `best_loss = np.mean([calc_loss(loss_op, model(Variable(xv)), Variable(yv)).data.cpu().numpy() for xv, yv in val_gen])`

@seongpyoHong There are opinions about several DB types. 1. The `bucket id` is hashed [here](https://github.com/graykode/matorage/blob/master/matorage/data/config.py#L129) and used as a string type. Therefore, it should be expressed as a string rather...