def test_base_model_inheritance_ver1(clean_config):
x = tf.ones([32, 64, 64, 3], dtype=tf.float32)
m1 = Conv2D('conv', 32, 3)
m2 = Dense('dense', 160)
m3 = Stack([m1, m2])
res1 = m3(x)
assert shape(res1) == [32, 64, 64, 160]
def _short_cut(self, name):
conv2d_ins = Conv2D(name=name,
filters=self.config[self.KEYS.CONFIG.FILTERS],
kernel_size=self.config[self.KEYS.CONFIG.KERNEL_SIZE])
return Stack(name='short_cut',
models=[conv2d_ins,
self._inference_blk(self.config[self.KEYS.CONFIG.NUM_INFER_BLK],
ResidualBlock(name='res_blk',
filters=self.config[self.KEYS.CONFIG.FILTERS]))])
def kernel(self, inputs):
x = inputs[self.KEYS.TENSOR.HITS]
x = flatten(x)
m = identity
models = []
for i in range(3):
models += [Dense(self.config(self.KEYS.CONFIG.NB_UNITS)
[i], info='dense_{}'.format(i)),
ReLU,
DropOut()]
# models.append(DropOut())
models.append(
Dense(self.config(self.KEYS.CONFIG.MAX_NB_HITS), info='dense_end'))
seq = self.graphs.get('seq', Stack(info='stack', models=models))
return seq(x)
def inference_blk(num_layer, layer):
l = []
for _ in range(num_layer):
l.append(layer)
return Stack(name='stack', models=l)
def _inference_blk(self, num_inference_blocks, layer):
l = []
for _ in range(num_inference_blocks):
l.append(layer)
return Stack(name='inference_blk', models=l)