diff --git a/bob/learn/em/mixture/gmm.py b/bob/learn/em/mixture/gmm.py
index e3f72ff7987c0e784359b9178f8d5b95873596cf..d64b6f190cdc387fca2fa8fbc17b5f2e82cb3f5b 100644
--- a/bob/learn/em/mixture/gmm.py
+++ b/bob/learn/em/mixture/gmm.py
@@ -480,9 +480,9 @@ class GMMMachine(BaseEstimator):
ubm=ubm,
weights=hdf5["m_weights"][()],
)
- self.means = g_means
- self.variances = g_variances
- self.variance_thresholds = g_variance_thresholds
+ self.means = np.array(g_means)
+ self.variances = np.array(g_variances)
+ self.variance_thresholds = np.array(g_variance_thresholds)
return self
def save(self, hdf5):
@@ -717,11 +717,13 @@ class GMMMachine(BaseEstimator):
# Note: Uses the stats from before m_step, leading to an additional m_step
# (which is not bad because it will always converge)
average_output = stats.log_likelihood / stats.t
+ logger.debug(f"average output = {average_output}")
if step > 1:
convergence_value = abs(
(average_output_previous - average_output) / average_output_previous
)
+ logger.debug(f"convergence val = {convergence_value}")
# Terminates if converged (and likelihood computation is set)
if (
diff --git a/bob/learn/em/test/test_gmm.py b/bob/learn/em/test/test_gmm.py
index 3f8dc37d7a51edb1f9854005c3d837bff9aa9f68..40d535a9fcaa1dba3d144769ce2be4618ace0ed1 100644
--- a/bob/learn/em/test/test_gmm.py
+++ b/bob/learn/em/test/test_gmm.py
@@ -542,3 +542,261 @@ def test_map_transformer():
sum_pxx=np.array([[2, 2, 5], [128, 145, 145]], dtype=float),
)
assert stats.is_similar_to(expected_stats)
+
+
+## Tests from `test_em.py`
+
+def loadGMM():
+ gmm = GMMMachine(2)
+
+ gmm.weights = bob.io.base.load(datafile("gmm.init_weights.hdf5", __name__, path="../data/"))
+ gmm.means = bob.io.base.load(datafile("gmm.init_means.hdf5", __name__, path="../data/"))
+ gmm.variances = bob.io.base.load(datafile("gmm.init_variances.hdf5", __name__, path="../data/"))
+
+ return gmm
+
+def equals(x, y, epsilon):
+ return (abs(x - y) < epsilon).all()
+
+def test_gmm_ML_1():
+ """Trains a GMMMachine with ML_GMMTrainer"""
+ ar = bob.io.base.load(datafile("faithful.torch3_f64.hdf5", __name__, path="../data/"))
+ gmm = loadGMM()
+
+ # test rng handling
+ gmm.convergence_threshold = 0.001
+ gmm.update_means = True
+ gmm.update_variances = True
+ gmm.update_weights = True
+ gmm.random_state = np.random.RandomState(seed=12345)
+ gmm = gmm.fit(ar)
+
+ gmm = loadGMM()
+ gmm.convergence_threshold = 0.001
+ gmm.update_means = True
+ gmm.update_variances = True
+ gmm.update_weights = True
+ gmm = gmm.fit(ar)
+
+ #config = HDF5File(datafile("gmm_ML.hdf5", __name__), "w")
+ #gmm.save(config)
+
+ gmm_ref = GMMMachine.from_hdf5(HDF5File(datafile("gmm_ML.hdf5", __name__, path="../data"), "r")) # TODO update the ref file(s)
+ gmm_ref_32bit_debug = GMMMachine.from_hdf5(HDF5File(datafile("gmm_ML_32bit_debug.hdf5", __name__, path="../data/"), "r"))
+ gmm_ref_32bit_release = GMMMachine.from_hdf5(HDF5File(datafile("gmm_ML_32bit_release.hdf5", __name__, path="../data/"), "r"))
+ assert (gmm == gmm_ref) # or (gmm == gmm_ref_32bit_release) or (gmm == gmm_ref_32bit_debug)
+
+
+def test_gmm_ML_2():
+ """Trains a GMMMachine with ML_GMMTrainer; compares to an old reference"""
+ ar = bob.io.base.load(datafile("dataNormalized.hdf5", __name__, path="../data/"))
+
+ # Initialize GMMMachine
+ gmm = GMMMachine(5, 45)
+ gmm.means = bob.io.base.load(datafile("meansAfterKMeans.hdf5", __name__, path="../data/")).astype("float64")
+ gmm.variances = bob.io.base.load(datafile("variancesAfterKMeans.hdf5", __name__, path="../data/")).astype("float64")
+ gmm.weights = np.exp(bob.io.base.load(datafile("weightsAfterKMeans.hdf5", __name__, path="../data/")).astype("float64"))
+
+ threshold = 0.001
+ gmm.variance_thresholds = threshold
+
+ # Initialize ML Trainer
+ gmm.mean_var_update_threshold = 0.001
+ gmm.max_fitting_steps = 26
+ gmm.convergence_threshold = 0.00001
+ gmm.update_means = True
+ gmm.update_variances = True
+ gmm.update_weights = True
+
+ # Run ML
+ gmm = gmm.fit(ar)
+ # Test results
+ # Load torch3vision reference
+ meansML_ref = bob.io.base.load(datafile("meansAfterML.hdf5", __name__, path="../data/"))
+ variancesML_ref = bob.io.base.load(datafile("variancesAfterML.hdf5", __name__, path="../data/"))
+ weightsML_ref = bob.io.base.load(datafile("weightsAfterML.hdf5", __name__, path="../data/"))
+
+ # Compare to current results
+ np.testing.assert_allclose(gmm.means, meansML_ref, rtol=3e-3)
+ np.testing.assert_allclose(gmm.variances, variancesML_ref, rtol=3e-3)
+ np.testing.assert_allclose(gmm.weights, weightsML_ref, rtol=1e-4)
+
+
+def test_gmm_ML_parallel():
+ """Trains a GMMMachine with ML_GMMTrainer; compares to an old reference"""
+
+ ar = da.array(bob.io.base.load(datafile("dataNormalized.hdf5", __name__, path="../data/")))
+
+ # Initialize GMMMachine
+ gmm = GMMMachine(5, 45)
+ gmm.means = bob.io.base.load(datafile("meansAfterKMeans.hdf5", __name__, path="../data/")).astype("float64")
+ gmm.variances = bob.io.base.load(datafile("variancesAfterKMeans.hdf5", __name__, path="../data/")).astype("float64")
+ gmm.weights = np.exp(bob.io.base.load(datafile("weightsAfterKMeans.hdf5", __name__, path="../data/")).astype("float64"))
+
+ threshold = 0.001
+ gmm.variance_thresholds = threshold
+
+ # Initialize ML Trainer
+ gmm.mean_var_update_threshold = 0.001
+ gmm.max_fitting_steps = 25
+ gmm.convergence_threshold = 0.00001
+ gmm.update_means = True
+ gmm.update_variances = True
+ gmm.update_weights = True
+
+ # Run ML
+ gmm.fit(ar)
+
+ # Test results
+ # Load torch3vision reference
+ meansML_ref = bob.io.base.load(datafile("meansAfterML.hdf5", __name__, path="../data/"))
+ variancesML_ref = bob.io.base.load(datafile("variancesAfterML.hdf5", __name__, path="../data/"))
+ weightsML_ref = bob.io.base.load(datafile("weightsAfterML.hdf5", __name__, path="../data/"))
+
+ # Compare to current results
+ np.testing.assert_allclose(gmm.means, meansML_ref, rtol=3e-3)
+ np.testing.assert_allclose(gmm.variances, variancesML_ref, rtol=3e-3)
+ np.testing.assert_allclose(gmm.weights, weightsML_ref, rtol=1e-4)
+
+
+
+def test_gmm_MAP_1():
+ """Train a GMMMachine with MAP_GMMTrainer"""
+ ar = bob.io.base.load(datafile("faithful.torch3_f64.hdf5", __name__, path="../data/"))
+
+ # test with rng
+ gmmprior = GMMMachine.from_hdf5(HDF5File(datafile("gmm_ML.hdf5", __name__, path="../data/"), "r"))
+ gmm = GMMMachine.from_hdf5(HDF5File(datafile("gmm_ML.hdf5", __name__, path="../data/"), "r"), ubm = gmmprior)
+ gmm.update_means = True
+ gmm.update_variances = False
+ gmm.update_weights = False
+ rng = np.random.RandomState(seed=12345)
+ gmm.random_state = rng
+ gmm = gmm.fit(ar)
+
+ gmmprior = GMMMachine.from_hdf5(HDF5File(datafile("gmm_ML.hdf5", __name__, path="../data/"), "r"))
+ gmm = GMMMachine.from_hdf5(HDF5File(datafile("gmm_ML.hdf5", __name__, path="../data/"), "r"), ubm = gmmprior)
+ gmm.update_means = True
+ gmm.update_variances = False
+ gmm.update_weights = False
+
+ gmm = gmm.fit(ar)
+
+ gmm_ref = GMMMachine.from_hdf5(HDF5File(datafile("gmm_MAP.hdf5", __name__, path="../data/"), "r"))
+
+ np.testing.assert_allclose(gmm.means, gmm_ref.means, rtol=1e-3)
+ np.testing.assert_allclose(gmm.variances, gmm_ref.variances, rtol=1e-3)
+ np.testing.assert_allclose(gmm.weights, gmm_ref.weights, rtol=1e-3)
+
+
+def test_gmm_MAP_2():
+ """Train a GMMMachine with MAP_GMMTrainer and compare with matlab reference"""
+
+ data = bob.io.base.load(datafile("data.hdf5", __name__, path="../data/"))
+ data = data.reshape((-1, 1)) # make a 2D array out of it
+ means = bob.io.base.load(datafile("means.hdf5", __name__, path="../data/"))
+ variances = bob.io.base.load(datafile("variances.hdf5", __name__, path="../data/"))
+ weights = bob.io.base.load(datafile("weights.hdf5", __name__, path="../data/"))
+
+ gmm = GMMMachine(n_gaussians=2)
+ gmm.means = means
+ gmm.variances = variances
+ gmm.weights = weights
+
+ gmm_adapted = GMMMachine(
+ n_gaussians=2,
+ trainer="map",
+ ubm=gmm,
+ max_fitting_steps=1,
+ update_means=True,
+ update_variances=False,
+ update_weights=False,
+ mean_var_update_threshold=0.,
+ )
+ gmm_adapted.means = means
+ gmm_adapted.variances = variances
+ gmm_adapted.weights = weights
+
+ gmm = gmm.fit(data)
+
+ gmm_adapted = gmm_adapted.fit(data)
+
+ new_means = bob.io.base.load(datafile("new_adapted_mean.hdf5", __name__, path="../data/"))
+
+ # Compare to matlab reference
+ np.testing.assert_allclose(new_means[0,:], gmm_adapted.means[:,0], rtol=1e-4)
+ np.testing.assert_allclose(new_means[1,:], gmm_adapted.means[:,1], rtol=1e-4)
+
+
+def test_gmm_MAP_3():
+ """Train a GMMMachine with MAP_GMMTrainer; compares to old reference"""
+ ar = bob.io.base.load(datafile("dataforMAP.hdf5", __name__, path="../data/"))
+
+ # Initialize GMMMachine
+ n_gaussians = 5
+ prior_gmm = GMMMachine(n_gaussians)
+ prior_gmm.means = bob.io.base.load(datafile("meansAfterML.hdf5", __name__, path="../data/"))
+ prior_gmm.variances = bob.io.base.load(datafile("variancesAfterML.hdf5", __name__, path="../data/"))
+ prior_gmm.weights = bob.io.base.load(datafile("weightsAfterML.hdf5", __name__, path="../data/"))
+
+ threshold = 0.001
+ prior_gmm.variance_thresholds = threshold
+
+ # Initialize MAP Trainer
+ accuracy = 0.00001
+
+ gmm = GMMMachine(
+ n_gaussians,
+ trainer="map",
+ ubm=prior_gmm,
+ convergence_threshold=threshold,
+ max_fitting_steps=1,
+ update_means=True,
+ update_variances=False,
+ update_weights=False,
+ mean_var_update_threshold=accuracy
+ )
+ gmm.variance_thresholds = threshold
+
+ # Train
+ gmm = gmm.fit(ar)
+
+ # Test results
+ # Load torch3vision reference
+ meansMAP_ref = bob.io.base.load(datafile("meansAfterMAP.hdf5", __name__, path="../data/"))
+ variancesMAP_ref = bob.io.base.load(datafile("variancesAfterMAP.hdf5", __name__, path="../data/"))
+ weightsMAP_ref = bob.io.base.load(datafile("weightsAfterMAP.hdf5", __name__, path="../data/"))
+
+ # Compare to current results
+ # Gaps are quite large. This might be explained by the fact that there is no
+ # adaptation of a given Gaussian in torch3 when the corresponding responsibilities
+ # are below the responsibilities threshold
+ np.testing.assert_allclose(gmm.means, meansMAP_ref, rtol=2e-1)
+ np.testing.assert_allclose(gmm.variances, variancesMAP_ref, rtol=1e-4)
+ np.testing.assert_allclose(gmm.weights, weightsMAP_ref, rtol=1e-4)
+
+
+def test_gmm_test():
+
+ # Tests a GMMMachine by computing scores against a model and compare to
+ # an old reference
+
+ ar = bob.io.base.load(datafile("dataforMAP.hdf5", __name__, path="../data/"))
+
+ # Initialize GMMMachine
+ n_gaussians = 5
+ gmm = GMMMachine(n_gaussians)
+ gmm.means = bob.io.base.load(datafile("meansAfterML.hdf5", __name__, path="../data/"))
+ gmm.variances = bob.io.base.load(datafile("variancesAfterML.hdf5", __name__, path="../data/"))
+ gmm.weights = bob.io.base.load(datafile("weightsAfterML.hdf5", __name__, path="../data/"))
+
+ threshold = 0.001
+ gmm.variance_thresholds = threshold
+
+ # Test against the model
+ score_mean_ref = -1.50379e+06
+ score = gmm.log_likelihood(ar).sum()
+ score /= len(ar)
+
+ # Compare current results to torch3vision
+ assert abs(score-score_mean_ref)/score_mean_ref<1e-4