XingWu_UCAS

@jeffra @tjruwase

> @1024er, apologies for the delay. Will take a closer look asap. Thank you so much ~ Is there any progress please? ：）

> @1024er, apologies I have not had much time to explore this. Hi，do you have time to take a look recently ? Thank you @tjruwase

> Hi @1024er , have you ever encountered the problem that DeepSpeed will stall at the end of the first epoch (please refer to [this issue](https://github.com/microsoft/DeepSpeedExamples/issues/122))? > > If so,...

> > > Hi @1024er , have you ever encountered the problem that DeepSpeed will stall at the end of the first epoch (please refer to [this issue](https://github.com/microsoft/DeepSpeedExamples/issues/122))? > >...

> https://github.com/PaddlePaddle/models/tree/c35c10a8fdd69039cc79c76f19cee87d04bc400b/PaddleNLP/PaddleTextGEN/seq2seq 好的，我尽快修复这个问题

> @1024er @Superjomn @reyoung 你们好，我想问一下，该模型中，计算的ppl值，在训练集，验证集和测试集都有计算的这个值，这是perplexity的计算吗？但网上查到的ppl公式不是根据ppl = np.exp(total_loss / word_count)计算的呀，所以这里的ppl是什么度量的计算呢？请求解惑，感谢。具体相关代码如下： > > ``` > def eval(data, epoch_id=0): > eval_data_iter = reader.get_data_iter(data, batch_size, mode='eval') > total_loss = 0.0 > word_count = 0.0...

@Victorwz

Sorry for not being able to respond in time due to busy work. I have switch the code to a new branch, which is more neat and easy to follow.

Sorry for not being able to respond in time due to busy work. I think it may happens, but we suggest to evaluate the model on validation data at the...