def m_step(self, expectations, datas, inputs, masks, tags, **kwargs):
x = np.concatenate(datas)
weights = np.concatenate([Ez for Ez, _, _ in expectations])
for k in range(self.K):
ps = np.clip(np.average(x, axis=0, weights=weights[:, k]), 1e-3,
1 - 1e-3)
self.logit_ps[k] = logit(ps)
def initialize(self, datas, inputs=None, masks=None, tags=None):
# Initialize with KMeans
from sklearn.cluster import KMeans
data = np.concatenate(datas)
km = KMeans(self.K).fit(data)
ps = np.clip(km.cluster_centers_, 1e-3, 1 - 1e-3)
self.logit_ps = logit(ps)