I’m training the weights of an embedding layer for POS tags (tagging done by nltk), and then appending this to my word embedding features before inputting the combined vector into an LSTM. Has anyone seen or dealt with the following issue before when working with embedding weight training?
File "/home/software/anaconda3/lib/python3.6/site-packages/torch/nn/_functions/dropout.py", line 39, in forward
ctx.noise.bernoulli_(1 - ctx.p).div_(1 - ctx.p)
RuntimeError: Creating MTGP constants failed. at /opt/conda/conda-bld/pytorch_1518244507981/work/torch/lib/THC/THCTensorRandom.cu:33```