search

malllabiisc / CompGCN

ICLR 2020: Composition-Based Multi-Relational Graph Convolutional Networks
Apache License 2.0
597 stars 107 forks source link

3 questions or suggestions #16

Closed PolarisRisingWar closed 4 years ago

PolarisRisingWar commented 4 years ago
  1. run.py: error: unrecognized arguments: -init_embed 200 After using "python run.py -score_func transe -opn sub -gamma 9 -hid_drop 0.1 -init_embed 200 " this sentence in terminal, I find that it occurred an error for me said "python run.py -score_func transe -opn sub -gamma 9 -hid_drop 0.1 -init_embed 200 ". I find that in "run.py" there isn't any "init_embed" code. Is that in fact a "-embed_dim" attribute or other attribute?
  2. I'm on Windows, so I have to change "'%H:%M:%S'" to "'%H%M%S' or so on. Coz Windows can't support ":" in the name of files.
  3. Honestly it's the first time for me to run a deep learning model... And it said my virtual memory is not enough(OSError: [WinError 1455]). I don't know how to deal with it. Maybe I have to buy a new computer? So I wonder how much memory do I finally need?
svjan5 commented 4 years ago

Hi @PolarisRisingWar,

  1. Thanks for pointing out the issue. It should have been init_dim instead of init_embed. I have made the correction in the Readme.
  2. We have tested the code only on Linux so not sure about its compatibility with windows. If it doesn't work then I would recommend using Linux OS only (dual boot / virtualbox).
  3. Yes, the code involves running CompGCN operation on the entire input knowledge graph so it might require some memory. But it works on an 11gb GPU. We used this for getting all our results.