Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

This code acc != source paper acc #14

Open
busishu opened this issue Apr 14, 2018 · 1 comment
Open

This code acc != source paper acc #14

busishu opened this issue Apr 14, 2018 · 1 comment

Comments

@busishu
Copy link

busishu commented Apr 14, 2018

Thx your code. But when i use this code in restaurant dataset, i run 100 epochs and got 70% acc.
Here is my hyper-parameters:
'batch_size': 1,
'edim': 300,
'init_hid': 0.1,
'init_lr': 0.01,
'init_std': 0.01,
'lindim': 100,
'max_grad_norm': 100,
'mem_size': 83,
'nepoch': 320,
'nhop': 9,
'nwords': 3066,
'pad_idx': 0,
'pretrain_embeddings': 'glove.840B.300d.txt',
'show': True,

I know something is wrong, but i cant fix it.
For example, increase batch_size but i got less than 20% accuracy. And i change the number of hops but nothing happen.

@amanchuan
Copy link

Hi, busishu. my accuracy is also less than 20%.... can you understand the meaning of "lindim" parameter? I don't understand the what's meaning of "linear part of the state".

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants