Initialisation of weights of the neural networks

cics-nd / rans-uncertainty

Uncertainty Quantification of RANS Data-Driven Turbulence Modeling

https://doi.org/10.1016/j.jcp.2019.01.021

MIT License

55 stars 31 forks source link

Initialisation of weights of the neural networks #14

Open deepakdalakoti opened 3 years ago

deepakdalakoti commented 3 years ago

Hi,

I think the weights of the neural network should be initialised by sampling from the prior distribution. In the current implementation, they are samples from a standard normal distribution. I think this will help in convergence.

Regards Deepak