yuyue9284

> @yuyue9284 Which Volcano version are you using? And could you post your mpi-operator deployment? > > Now, volcano v1.0.0 did change the PodGroup CRD APIGroup to `volcano.sh`, so you...

> Hi @yuyue9284, thank you for your feedback. > > Please provide your detailed code to better solve your problem. Hi @Wzb123456789, this is how we used it in our...

您好，就是cost的定义部分是 ``` cost = tf.contrib.seq2seq.sequence_loss( training_logits, targets, masks) ``` 然后在计算validatoin loss 部分还是调用的cost，虽然输入数据时用的是validation data，但是由于上面cost计算用的是training_logits，最后应该还是用的training decoder的输出，而不是用的predicting decoder。 ``` # 计算validation loss validation_loss = sess.run( [cost], {input_data: valid_sources_batch, targets: valid_targets_batch, lr: learning_rate, target_sequence_length:...

了解了，谢谢，这个系列太好了，对学习很有帮助👍！