How to use gpu to run trainer? about matchzoo-py

[Chriskuei]

The latest version of MatchZoo-py (v1.1.1) has solved this issue

When I run the drmm code in tutorials, there is an error says:
Traceback (most recent call last):
File "pydrmm.py", line 64, in
callback=padding_callback
TypeError: init() got an unexpected keyword argument 'batch_size'
Why it occured?

DRMM tutorial has not been updated yet, you can refer to DRMMTKS tutorial

[Chriskuei]

Please check if CUDA is currently available by using torch.cuda.is_available().

[czt616]

torch.cuda.is_available() returned true. But the speed seems the same as running on CPU.

[lianrzh]

@czt616 Try to change the parameter 'device' by setting device=torch.device('cpu') to check the speed of cpu. The default value of 'device' is torch.device('cuda') when CUDA is available.

[PathricLee]

I have the same problems, when torch.cuda.is_available() returned true, but the speed is still now. the gpu rate of utilization is about 10% and most of time is 0%.

[Zhiyu-Chen]

It may result from this issue: #109

[Chriskuei]

The latest version of MatchZoo-py (v1.1.1) has solved this issue

[czt616]

The latest version of MatchZoo-py (v1.1.1) has solved this issue

When I run the drmm code in tutorials, there is an error says:
Traceback (most recent call last):
File "pydrmm.py", line 64, in
callback=padding_callback
TypeError: init() got an unexpected keyword argument 'batch_size'
Why it occured?