search

Smerity / sha-rnn

Single Headed Attention RNN - "Stop thinking with your head"
1.18k stars 133 forks source link

Could you share pretrained model weights? #2

Open vhargitai opened 4 years ago

vhargitai commented 4 years ago

Hi @Smerity , could you share the pretrained SHA-RNN weights from your WikiText103 experiments? I'd like to do some fine-tuning experiments with it for text classification. (It would be a huge help, as I only have access to an oven for limited time. :)) Thank you!

PiotrCzapla commented 4 years ago

Steven It would be useful to have an LSTM based network shown on wikitext103 benchmark on sotabench. If you don't have the pre-trained weights anymore I can run the training but it would be good to host it here in this repo.
https://sotabench.com/benchmarks/language-modelling-on-wikitext-103