check the unit test VGG11.test_density_fit()

[charlesmartin14]

Check VGG11.test_density_fit

Thre may be something wrong with the MPFits

details = self.watcher.analyze(layers = [self.fc1_layer], pool=False, randomize=False, plot=False, mp_fit=True) There is something wrong with this test because the sigma is way too low for layer fc1

A good sigma should be around 1.0, but this gives sigma~0.7

See also issue: #190

[charlesmartin14]

Referring to #210, what i suspect is happening here is that the MP fit is overfitting near zero

[charlesmartin14]

What I suspect is happening here is that we need to consider

• the rescaling of the density
• removing the eigenvalues below lambda_min