Alexandre Pinto comments

Repositories
Issues
Comments

Results 3 comments of


                                            Alexandre Pinto

Using XLNetModel class for inference

Don't forget that the get_pooled_out method uses a linear projection at the end of the pooling process, by default. ```python get_pooled_out(summary_type, use_summ_proj=True) ``` This adds an additional Dense layer. If...

Using XLNetModel class for inference

You can set the seed to a fixed value in order to avoid the randomness. The initializer is controled by the `FLAGS.init` flag, which can be 'uniform' or 'normal'. That...

different runs on same data produce slightly different outputs

You can turn it off mannuay: ``` xlnet_config.dropout = 0.0 xlnet_config.dropatt = 0.0 ```