-
## 🐛 Bug
I can train Transformers but not Fully convolutional or LSTMs models (e.g.: `fconv,fconv_iwslt_de_en, fconv_wmt_en_de, lstm, lstm_luong_wmt_en_de,`...) because gradients are inconsistent b…
-
Hi!
Recently I stumbled across your repo and wmt models. They showed pretty good results on my data out-of-the-box (I uploaded them via HuggingFace) but I failed to find any information about how t…
-
### Author Pages
https://aclanthology.org/people/y/yifan-peng/
### Type of Author Metadata Correction
- [X] The author page wrongly conflates different people with the same name.
- [ ] This author …
pyf98 updated
4 months ago
-
## 🐛 Bug
Multi-head attention module throwing error: "RuntimeError: view size is not compatible with input tensor's size and stride (at least one dimension spans across two contiguous subspaces). Use…
-
### 論文へのリンク
[[arXiv:2004.11886] Lite Transformer with Long-Short Range Attention](https://arxiv.org/abs/2004.11886)
### 著者・所属機関
Zhanghao Wu, Zhijian Liu, Ji Lin, Yujun Lin, Song Han
- MIT
…
-
Hi, thanks for the great work ! When trying to replicate the results on `iwslt14.de-en` you reported, I encountered the following problems. Your kind help will be much appreciated !
## data preproc…
-
Hi:
Just want to know How to replicate the result you mentioned on README, `The model reaches 20 BLEU on testing dataset, after training for only 2 epochs`.
I simple used your setup to train my…
-
- https://www.aclweb.org/anthology/W18-6319/
- 2018
機械翻訳の分野では、その主要な評価指標であるBLEUスコアの報告に一貫性がないため、あまり認識されていない問題に直面しています。
人々はBLEUスコアを「The」と呼んでいますが、BLEUは実際にはパラメータ化された指標であり、その値はパラメータの変更によって大きく変化します。
その…
e4exp updated
3 years ago
-
Hi there,
I am very interested in your work, but my computational resources are limited, so I would like to try a smaller dataset, such as IWSLT14, with settings N=6 and M=3 or N=12, 18 and M = 6. …
-
I am playing with the MMA-hard model to replicate WMT15 DE-EN experiments reported in the paper and my question is about preprocessing and postprocessing data. The paper says that:
> For each data…