thunlp / NRE

Neural Relation Extraction, including CNN, PCNN, CNN+ATT, PCNN+ATT
MIT License
810 stars 309 forks source link

Segmentation Fault #4

Closed SharmisthaJat closed 8 years ago

SharmisthaJat commented 8 years ago

Hi,

Thanks for sharing your code. I am running this on a linux server with 60GB free memory. Compile works fine, but ./train leads to a segmentation fault.

[s CNN+ATT]$ ls init.h log.txt makefile out test.cpp test.h train.cpp [s CNN+ATT]$ make g++ train.cpp -o train -O2 -lpthread g++ test.cpp -o test -O2 -lpthread [sharmistha@momo CNN+ATT]$ ./train Init Begin.

Segmentation fault (core dumped)

All the training files are in the correct directory. Please let me know how to resolve this issue.

Thanks

Mrlyk423 commented 8 years ago

Please sure that you have unzipped the data.zip in data folder.

SharmisthaJat commented 8 years ago

data is unzipped in the data folder. Following are the contents,

[s]$ ls data/ att.test data entity2id.txt test.txt vec.bin att.train data.zip relation2id.txt train.txt vector1.txt

Do I need anything else in the data folder to make this work ?

Mrlyk423 commented 8 years ago

Please follow the original data location after unzipping. The files expect "vec.bin" need to be placed in the folder "RE".

SharmisthaJat commented 8 years ago

Thanks, Moving vec.bin to RE works.

SharmisthaJat commented 8 years ago

I had a minor question. The MultiR dataset contained about 120k train examples and 180k test examples in the protocol buffer format. While you have almost 522611train examples. Did you preprocess the data again from the riedel dataset ? and if so what was the process like.