MatchZoo-py

MatchZoo-py copied to clipboard

### Describe the bug Hi, the implement about the MAP is not right. In my opinion, the denominator of every AP should be the number of relevant documents counted from...

wuchen95

Is there any tutorial for using the pretrained model for predicting new sentence pair?

1

I have trained the Esim and Bert model. However, I can not found any code for predicting new sentence pair? Is anyone can help?

anbo724

In the attention.py file of modules, line 53 is the process of calculating the attention weight in the following picture. The value of mask filling should be -1e7, right?

HuipengXu

DRCN model

1

I hope the author can publish a version of the drcn model

gldashen

RankHingeLoss代码逻辑不合理

2

### 描述 RankHingeLoss的forward方法计算逻辑是：先计算负样本的cosine均值，再调用margin_ranking_loss计算loss。这样是不合理的，因为均值的margin_ranking_loss不等于margin_ranking_loss的均值。我尝试更改后，在模型效果上有很大提升。 ### fix后的参考代码 def forward(self, y_pred: torch.Tensor, y_true: torch.Tensor): """ Calculate rank hinge loss. :param y_pred: Predicted result. :param y_true: Label. :return: Hinge loss computed by user-defined margin....

wenjunyang

### Describe the bug I cloned this repo by: `git clone https://github.com/NTMC-Community/MatchZoo-py.git` `cd MatchZoo-py` `python setup.py install` then I went to tutorials dir: `cd tutorials/ranking` To run .ipynb file in...

surzia

Split workload when use multiprocessing for loading data.

1

#150

Albert-Ma

dalek-who

`trainloader = mz.dataloader.DataLoader( dataset=trainset, stage='train', num_workers=30, callback=padding_callback )`` Result:  It shows that model will train `num_worker` time, 30 in this experiment, in one epoch! A Possible Solution: splitting workload...

Albert-Ma

DSSM tutorial code: RuntimeError Expected object of scalar type Float but got scalar type Double

2

### Describe the bug When I run the tutorial code for DSSM, the following error appears: **RuntimeError: Expected object of scalar type Float but got scalar type Double for argument...

katestfla