def test_symbol_saveload():
sym = models.mlp2()
fname = 'tmp_sym.json'
sym.save(fname)
data2 = mx.symbol.load(fname)
# save because of order
assert sym.tojson() == data2.tojson()
os.remove(fname)
def test_mlp2_infer_shape():
# Build MLP
out = models.mlp2()
# infer shape
data_shape = (100, 100)
arg_shapes, out_shapes, aux_shapes = out.infer_shape(data=data_shape)
arg_shape_dict = dict(zip(out.list_arguments(), arg_shapes))
assert len(out_shapes) == 1
assert out_shapes[0] == (100, 10)
true_shapes = {'fc2_bias': (10,),
'fc2_weight' : (10, 1000),
'fc1_bias' : (1000,),
'fc1_weight' : (1000,100) }
for k, v in true_shapes.items():
assert arg_shape_dict[k] == v
def test_symbol_basic():
print "======================test_symbol_basic========================"
"""models.mlp2 会产生一个含有两个隐藏层的mlp
list_arguments: ['data', 'fc1_weight', 'fc1_bias', 'fc2_weight', 'fc2_bias']
list_outputs: ['fc2_output']"""
mlist = []
mlist.append(models.mlp2())
for m in mlist:
m.list_arguments()
print 'list_argumens'
print m.list_arguments()
m.list_outputs()
print 'list_outputs'
print m.list_outputs()
print "======================test_symbol_basic========================"
def test_symbol_basic():
mlist = []
mlist.append(models.mlp2())
for m in mlist:
m.list_arguments()
m.list_outputs()
def test_symbol_pickle():
mlist = [models.mlp2(), models.conv()]
data = pkl.dumps(mlist)
mlist2 = pkl.loads(data)
for x, y in zip(mlist, mlist2):
assert x.tojson() == y.tojson()
def test_mlp2_infer_error():
# Test shape inconsistent case
out = models.mlp2()
weight_shape= (1, 100)
data_shape = (100, 100)
arg_shapes, out_shapes, aux_shapes = out.infer_shape(data=data_shape, fc1_weight=weight_shape)
def test_symbol_basic():
mlist = []
mlist.append(models.mlp2())
for m in mlist:
m.list_arguments()
m.list_outputs()
def test_symbol_pickle():
mlist = [models.mlp2(), models.conv()]
data = pkl.dumps(mlist)
mlist2 = pkl.loads(data)
for x, y in zip(mlist, mlist2):
assert x.tojson() == y.tojson()
def test_mlp2_infer_error():
# Test shape inconsistent case
out = models.mlp2()
weight_shape= (1, 100)
data_shape = (100, 100)
arg_shapes, out_shapes, aux_shapes = out.infer_shape(data=data_shape, fc1_weight=weight_shape)
