Xiaochuan Zhi
Xiaochuan Zhi
Hello, I need to reproduce the results on a subset of your dataset and I met some problems including `pid killed` in parsing, `ascii error` in `create_*_1.py` and `key error`...
I installed graphvite in an ubuntu18-cudnn7-cuda10.1-python3.7 version docker. conda version is `conda 4.8.3` `conda list graphvite` got the following result: > \# packages in environment at /usr/local/anaonda3: > \# >...
in wiki_utils.py three files are mentioned, but can not be found. ``` feat_LS = np.load('../data/wiki/Extracted_Features/lex.npy') feat_read = np.load('../data/wiki/Extracted_Features/read.npy') feat_numeric = np.load('../data/wiki/Extracted_Features/numeric.npy') ```
cost:nan
I used my own data to train, and cost nan occured. I checked the data, clipped the gradient, and reduced the learning rate, it still occured at the same 'batch_size*batch'...
python test.py --ws=4 --l2_reg=0.0004 --epoch=20 --max_len=40 --model_type=ABCNN3 --num_layers=2 --data_type=WebQA --classifier=LR ================================================== List of Variables: CNN-1/aW:0 CNN-1/conv/weights:0 CNN-1/conv/biases:0 CNN-2/aW:0 CNN-2/conv/weights:0 CNN-2/conv/biases:0 output-layer/FC/weights:0 output-layer/FC/biases:0 ================================================== ================================================== test data size: 10 2018-04-08 12:11:47.494259:...
info.dat explains the domain and range of each relation, which should be indispensable. LP datasets like amazon and LastFM own the 'info.dat', but PubMed does not. Did you miss it,...
If you release the valid metrics, we can compare it with simple-HGB on valid metrics locally, since we can not obtain the test results easily.
I used a 32g V100 to train RGCN on freebase dataset, but get the following error: > Using backend: pytorch > begin: 4194304 > Traceback (most recent call last): >...
Hi, I want to extend your code to implement a binary link prediction task, the train_X is the indexes of nodes in graph with the shape of `(train_size,2)`, each pair...