def test_log_softmax():
ndim = 2
shape = (SMALL_Y, LARGE_X)
axis = np.random.randint(0, ndim)
data = np.random.uniform(-2, 2, size=shape)
sym = mx.sym.log_softmax(axis=axis-ndim)
check_symbolic_forward(sym, [data], [np.log(np_softmax(data, axis=axis)+1e-20)])
def test_sigmoid():
def fsigmoid(a):
return np.divide(1.0, (1.0 + np.exp(-a)))
shape = (SMALL_Y, LARGE_X)
x = mx.symbol.Variable("x")
y = mx.sym.sigmoid(x)
xa = np.random.uniform(low=-1.0,high=1.0,size=shape)
ya = fsigmoid(xa)
check_symbolic_forward(y, [xa], [ya])
def test_relu():
def frelu(x):
return np.maximum(x, 0.0)
def frelu_grad(x):
return 1.0 * (x > 0.0)
shape = (SMALL_Y, LARGE_X)
x = mx.symbol.Variable("x")
y = mx.sym.relu(x)
xa = np.random.uniform(low=-1.0,high=1.0,size=shape)
eps = 1e-4
xa[abs(xa) < eps] = 1.0
ya = frelu(xa)
ga = frelu_grad(xa)
check_symbolic_forward(y, [xa], [ya])