You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Thanks a lot to the pybrain team for developing this amazing package!
I was trying to use the ExactNES optimization algorithm which efficiently computes the exact inverse of the Fisher Information Matrix. The ExactNES optimizer seems to terminate prematurely with the message: Numerical Instability. Stopping.
This is a simple test script for reproducing the error:
importnumpyasnpfrompybrain.optimizationimportExactNES# Set seed for repeatabilitynp.random.seed(0)
# Loss functiondefsphere(x):
returnnp.sum(x**2)
# Problem parametersdim=2x0=np.ones(dim)
# Initialize the optimizeroptim=ExactNES(sphere, x0)
optim.minimize=Trueoptim.verbose=Trueoptim.maxLearningSteps=1000# Trainx_best, best_val=optim.learn()
print("Best Value:", best_val)
I must also mention a couple of minor changes I had to make in the source code without which I was getting a type error (`TypeError: 'float' object cannot be interpreted as an integer'). The explicit changes are:
Thanks a lot to the pybrain team for developing this amazing package!
I was trying to use the ExactNES optimization algorithm which efficiently computes the exact inverse of the Fisher Information Matrix. The ExactNES optimizer seems to terminate prematurely with the message:
Numerical Instability. Stopping.
This is a simple test script for reproducing the error:
This gives the following output:
I must also mention a couple of minor changes I had to make in the source code without which I was getting a type error (`TypeError: 'float' object cannot be interpreted as an integer'). The explicit changes are:
xdim + xdim * (xdim + 1) / 2
toint(xdim + xdim * (xdim + 1) / 2)
inpybrain/pybrain/optimization/distributionbased/ves.py
Line 50 in dcdf32b
dim * (dim + 1) / 2
toint(dim * (dim + 1) / 2)
inpybrain/pybrain/utilities.py
Line 560 in dcdf32b
System Details: Ubuntu 16.04; running in a conda environment with Python 3.7.3.
Please let me know if any further details are necessary.
The text was updated successfully, but these errors were encountered: