Zhixing Tan

Results 14 comments of Zhixing Tan

预训练模型

你好，我们暂时还没有释放已训练好的模型的计划。

batch_size 10G GPU 单机最大能做到多少

这个值取决于模型大小以及GPU是否支持半精度。可以先设置一个值，然后观察训练时是否出现OOM，如有则继续调小。

pre-trained models?

The size of pre-trained models are too big (over 1GB). Unfortunately, we are unable to upload such big files to Google Drive due to some restrictions in China's networking. We...

[Question] Does dropout layer after word embedding improve neural machine translation?

We found this trick works well in practice. However, the original GroundHog implementation do not use this trick.