training config for wikitext103

[pichuang1984]

Thanks so much for open-source your code!

Will it be possible to provide training/config for wikitext-103 dataset?

Thanks

[Zhanghao Wu]

Thank you for asking! We just added the codes and configs in the new branch language-model with the pre-trained checkpoints.

[pichuang1984]

Thanks for providing the codes and configs. However it seems that the architecture transformer_lm_multibranch_v2_wiki103_small is not available in transformer_multibranch_v2.py?

[Zhanghao Wu]

Hi, the model is in the file lite-transformer/fairseq/models/transformer_lm_multibranch_v2.py in the language-model branch.

[pichuang1984]

Thanks, you are right, I was still looking at transformer_multibranch_v2.py