Anwen Hu

Nan problem during training on ontonotes data set

7

Hi Strubel, I ran "./bin/train_cnn.sh ./conf/ontonotes/bilstm-viterbi-doc.conf" for ontonotes data set. In the tenth iteration, there was a NAN problem, and a running warning that "RuntimeWarning: invalid value encountered in reduce...

how LSTM+Glove+IA encode articles?

8

@alasdairtran Hi, I have read your newly published paper. I'm curious about how LSTM+Glove+IA encodes articles? Does it encode each article at the sentence level or the word level?

Fintuning code on FUNSD and CORD

3

Thanks for your contribution for Document AI. There are scripts for fine-tuning RVL-CDIP and DUE-benchmark，but the one for FUNSD and CORD seems missing. Could you also share this script? Besides,...

RuntimeError: expected a non-empty list of Tensors

1

@furkanbiten Hi, I got a RuntimeError, when I ran training code where the 'caption_model' was set as 'adaatt'. The detail was shown in the attached image. (  ) (I...

scene.assets is None when load a scene

2

Hi, thanks for your wonderful work! When exploring this tool, I found that **scene.assets is empty when loading an existing scene**. I'm sure there are multiple objects in the scene...

Hi, @CurryYuan, thanks for your great work! I'm confused about why the teacher model only uses 2D features as inputs during training (line 137 in xtrans.py). As written in the...

RuntimeError: Expected object of type torch.cuda.FloatTensor but found type torch.FloatTensor for argument #4 'other'

1

Hi, when I evaluate a pretrained model named `att2in', I get a Runtime Error. I think there is nothing wrong with my environment (py2.7 torch0.4.1). Could you give some suggestion?...