def simulate(self, dt):
try:
engine.simulate(self, self.data, dt)
self.simulate_error = None
except Exception as error:
if self.simulate_error != error.message:
print("Unable to render, reason:")
print(error)
self.simulate_error = error.message
def test_small_batch_size(self):
tree = self.tree
model = self.model
board = self.board
simulate(tree, board, model, mcts_batch_size=1, original_player=1)
self.assertEqual(tree['subtree'][0]['count'], 1)
self.assertEqual(tree['subtree'][0]['value'], -1)
self.assertNotEqual(tree['subtree'][0]['subtree'], {})
self.assertEqual(tree['subtree'][1]['count'], 0)
self.assertEqual(tree['subtree'][1]['value'], 0)
self.assertEqual(tree['subtree'][1]['subtree'], {})
def test_leaf(self):
tree = self.tree
board = self.board
model = self.model
simulate(tree, board, model, mcts_batch_size=2, original_player=1)
self.assertEqual(tree['subtree'][0]['count'], 1)
self.assertEqual(tree['subtree'][1]['count'], 1)
self.assertEqual(tree['subtree'][0]['value'], -1)
self.assertEqual(tree['subtree'][1]['value'], -1)
self.assertEqual(tree['count'], 2)
self.assertEqual(tree['value'], -2)
self.assertEqual(tree['mean_value'], -1)
def test_nested_other_leaves(self):
model = self.model
board = self.board
tree = {
'count': 0,
'mean_value': 0,
'value': 0,
'parent': None,
'move': 1,
'subtree': {
0:{
'count': 0,
'p': .75,
'value': 0,
'mean_value': 0,
'move': 2,
'subtree': {}
},
1: {
'count': 0,
'p': .25,
'mean_value': 0,
'value': 0,
'move': 2,
'subtree': {
0: {
'count': 0,
'p': 1,
'mean_value': 0,
'value': 0,
'move': 3,
'subtree': {},
},
2: {
'count': 0,
'p': 0,
'mean_value': 0,
'value': 0,
'move': 3,
'subtree': {},
}
}
},
2:{
'count': 0,
'p': 0,
'value': 0,
'mean_value': 0,
'move': 2,
'subtree': {}
},
}
}
tree['subtree'][0]['parent'] = tree
tree['subtree'][1]['parent'] = tree
tree['subtree'][1]['subtree'][0]['parent'] = tree['subtree'][1]
tree['subtree'][1]['subtree'][2]['parent'] = tree['subtree'][1]
simulate(tree, board, model, mcts_batch_size=2, original_player=1)
self.assertEqual(tree['subtree'][0]['count'], 1)
self.assertEqual(tree['subtree'][0]['value'], -1)
self.assertEqual(tree['subtree'][1]['value'], 1)
self.assertEqual(tree['subtree'][1]['count'], 1)
self.assertEqual(tree['subtree'][1]['subtree'][0]['count'], 1)
self.assertEqual(tree['subtree'][1]['subtree'][0]['value'], 1)
self.assertEqual(tree['subtree'][1]['subtree'][2]['count'], 0)
self.assertEqual(tree['count'], 2)
self.assertEqual(tree['mean_value'], 0)
self.assertEqual(tree['subtree'][2]['count'], 0)
self.assertEqual(tree['subtree'][2]['subtree'], {})
def test_model_evaluation_other_nested(self):
tree = {
'count': 0,
'mean_value': 0,
'value': 0,
'parent': None,
'move': 1,
'subtree':{
0:{
'count': 0,
'p': 1,
'value': 0,
'mean_value': 0,
'move': 2,
'subtree': {},
},
1: {
'count': 0,
'p': 0,
'mean_value': 0,
'value': 0,
'move': 2,
'subtree': {
0: {
'count': 0,
'p': 0,
'mean_value': 0,
'value': 0,
'move': 3,
'subtree': {},
},
2: {
'count': 0,
'p': 1,
'mean_value': 0,
'value': 0,
'move': 3,
'subtree': {},
}
}
}
}
}
tree['subtree'][0]['parent'] = tree
tree['subtree'][1]['parent'] = tree
tree['subtree'][1]['subtree'][0]['parent'] = tree['subtree'][1]
tree['subtree'][1]['subtree'][2]['parent'] = tree['subtree'][1]
board = self.board
test_board1, player = game_init()
make_play(0, 0, test_board1)
test_board2, player = game_init()
make_play(1, 0, test_board2)
make_play(2, 0, test_board2)
class DummyModel(object):
def predict_on_batch(_, X):
size = conf['SIZE']
board1 = X[0].reshape(1, size, size, 17)
board2 = X[1].reshape(1, size, size, 17)
self.assertTrue(np.array_equal(board1, test_board1))
self.assertTrue(np.array_equal(board2, test_board2))
batch_size = X.shape[0]
policy = np.zeros((batch_size, size * size + 1), dtype=np.float32)
policy[:,0] = 1
value = np.zeros((batch_size, 1), dtype=np.float32)
value[:] = 1
return policy, value
model = DummyModel()
simulate(tree, board, model, mcts_batch_size=2, original_player=1)
def model():
inputs = engine.sample(
'input', dist.Normal(loc=np.array([0.]), scale=np.array([1.])))
outputs = engine.simulate('gn', inputs, obs=obs)
return inputs, outputs
def model():
inputs = engine.sample(
'input',
dist.Normal(loc=torch.tensor([0.]), scale=torch.tensor([1.])))
outputs = engine.simulate('gn', inputs, obs=obs)
return inputs, outputs
