Deep neural network for high level reasioning
The combinatorial model over sets of high-level representation is a crucial part of
our intelligent behavior thus, to achieve a more natural and complex system in NLU, a higher level of representation is mandatory. The conceptual base theory has an initial premise
that the bases of any natural language (NL) is conceptual base a representation that is not necessarily fixed or tied to a single or group of words but rather concepts and conceptualization is regarded as the relationship between these representations.
The following libraries meke the dependencies for the model
- Python3
- PyTorch
- TensorFlow
- tensorboardX
- matplotlib
The model works for
Example: The default program runs on multiple gpu's The main.py contains list of GPU's to run the program on.
python main.py # Run on multiple GPU's
However if you want to run on CPU use the use_cuda option
python main.py --use_cuda false # Run on CPU
Note: Eventhough the model can run on cpu, it highly recomanded to use Multiple GPU to train the model.
UESTC