Merge pull request #154 from hyperion-ml/persephone-refactor

debug gmm

Author: jesus-villalba

hyperion/np/clustering/kmeans.py

@@ -188,14 +188,14 @@ def _compute_centroids(self, x, index):
         for k in range(self.num_clusters):
             r = index == k
             if np.sum(r) > 0:
-                mu[k] = np.mean(x[index == k], axis=0)
+                mu[k] = np.mean(x[r], axis=0)
         return mu
 
     @staticmethod
     def _compute_centroid(x, index, k):
         r = index == k
         if np.sum(r) > 0:
-            return np.mean(x[index == k], axis=0)
+            return np.mean(x[r], axis=0)
         else:
             return None
 

hyperion/np/pdfs/mixtures/exp_family_mixture.py

@@ -2,6 +2,7 @@
  Copyright 2018 Johns Hopkins University  (Author: Jesus Villalba)
  Apache 2.0  (http://www.apache.org/licenses/LICENSE-2.0)
 """
+
 import logging
 
 import numpy as np
@@ -104,6 +105,8 @@ def fit(
                 )
                 elbo_val[epoch] = self.elbo(None, N=N, u_x=u_x, log_h=log_h_val)
 
+            print(self.log_prob(x, mode="nat"), self.log_prob(x, mode="std"))
+
         if x_val is None:
             return elbo, elbo / x.shape[0]
         else:

hyperion/np/pdfs/mixtures/gmm_diag_cov.py

@@ -180,9 +180,11 @@ def Mstep(self, N, u_x):
 
         if self.update_Lambda:
             S = S / N[:, None] - self.mu**2
+            print("1", S)
             S_floor = self.var_floor * np.mean(S[N > self.min_N], axis=0)
             S_floor = np.maximum(S_floor, 1e-10)
             S = np.maximum(S, S_floor)
+            print("2", S)
             print(np.min(S))
             self.Lambda = 1 / S
             self._Sigma = S