Merge branch 'tf' into 'master'

Stop the TensorflowEval algorithm to load on imports

See merge request !17

bob/pad/voice/algorithm/TensorflowEval.py

@@ -36,8 +36,7 @@ class TensorflowEval(Algorithm):
         self.num_time_steps = input_shape[0]
         self.network_size = network_size
         self.data_std = None
-#        import ipdb
-#        ipdb.set_trace()
+
         features_length = input_shape[1]
         if normalization_file and os.path.exists(normalization_file):
             logger.info("Loading normalization file '%s' " % normalization_file)
@@ -240,6 +239,3 @@ class TensorflowEval(Algorithm):
         logger.debug("score() score %f", toscore)
         # return only real score
         return [toscore[0]]
-
-
-algorithm = TensorflowEval()

bob/pad/voice/config/algorithm/tensorfloweval.py

+from bob.pad.voice.algorithm import TensorflowEval
+
+algorithm = TensorflowEval()

setup.py

@@ -116,7 +116,7 @@ setup(
         ],
 
         'bob.pad.algorithm': [
-            'tensorflow  = bob.pad.voice.algorithm.TensorflowEval:algorithm',
+            'tensorflow  = bob.pad.voice.config.algorithm.tensorfloweval:algorithm',
             'dummy-algo  = bob.pad.voice.algorithm.dummy:algorithm',
             # compute scores based on different energy bands
             'logregr  = bob.pad.voice.algorithm.LogRegr:algorithm',