GMFTBY

我的意思是，如果保留了是不是效果会更好呢？

谢谢

好的，您在 8 * 8 棋盘上模型的训练运行大约了 2 天，请问你训练这个 8 * 8 的硬件配置是怎么样的，非常感谢

但是在调用的时候传入的应该是 leaf_value 而不是 -leaf_value 啊，update_recursive 函数中的负号的含义很明确，但是这里感觉需要传入的是 leaf_value ??希望可以解释一下，这里看的不是很懂 @junxiaosong

Also confused about the `entropies` in the loss function, can anyone make a little explanation ?

回复晚了非常抱歉，这一句是为了兼容encoder和decoder的

Hi, thanks for your attention on this repo. Compared with the results in the original DailyDialog paper, the BLEU-1/2 score are lower but it can also be found that the...

Typically, the checkpoint need to be saved when the lowest losses (ppl) are achieved. But during my experiments, I find the model' performance can be further improved by training more...

Thank you so much for your attention to this repo. 1. As for the GAN-based model, it will take me some time to implement it, which may take about one...

Sorry for the late response, I have been busy recently. After my experiments, I found that DSHRED-WA is the best one. But it also costs lots of time to converge....