-
Notifications
You must be signed in to change notification settings - Fork 0
/
train_spider.py
108 lines (81 loc) · 3.43 KB
/
train_spider.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
import numpy as np
import pandas as pd
import os, warnings
warnings.filterwarnings("ignore")
os.environ["TOKENIZERS_PARALLELISM"] = "false"
from main import generate_diag
from pytket.extensions.qiskit import AerBackend
from lambeq import AtomicType, IQPAnsatz, TketModel, QuantumTrainer, SPSAOptimizer, Dataset
def read_data(filename):
labels = []
df = pd.read_csv(filename, index_col=0)
for i in range(df.shape[0]):
l = df.iloc[i]['label']
t = int(l)
labels.append([t, 1-t])
return labels
train_labels = read_data('dataset_144/train_72.csv')
dev_labels = read_data('dataset_144/dev_36.csv')
test_labels = read_data('dataset_144/test_36.csv')
conf = 'spider'
train_diagrams = generate_diag(pd.read_csv('dataset_144/train_72.csv', index_col=0), conf)
dev_diagrams = generate_diag(pd.read_csv('dataset_144/dev_36.csv', index_col=0), conf)
test_diagrams = generate_diag(pd.read_csv('dataset_144/test_36.csv', index_col=0), conf)
ansatz = IQPAnsatz({AtomicType.NOUN: 1, AtomicType.SENTENCE: 1},n_layers=1, n_single_qubit_params=3)
train_circuits = [ansatz(diagram) for diagram in train_diagrams]
dev_circuits = [ansatz(diagram) for diagram in dev_diagrams]
test_circuits = [ansatz(diagram) for diagram in test_diagrams]
all_circuits = train_circuits+dev_circuits+test_circuits
backend = AerBackend()
backend_config = {
'backend': backend,
'compilation': backend.default_compilation_pass(2),
'shots': 8192
}
model = TketModel.from_diagrams(all_circuits, backend_config=backend_config)
loss = lambda y_hat, y: -np.sum(y * np.log(y_hat)) / len(y) # binary cross-entropy loss
acc = lambda y_hat, y: np.sum(np.round(y_hat) == y) / len(y) / 2 # half due to double-counting
BATCH_SIZE = 36
EPOCHS = 100
train_dataset = Dataset(
train_circuits,
train_labels,
batch_size=BATCH_SIZE)
val_dataset = Dataset(dev_circuits, dev_labels, shuffle=True)
runs = 20
for i in range(runs):
print('run: ', i)
SEED = np.random.randint(1000)
print(SEED)
trainer = QuantumTrainer(
model,
loss_function=loss,
epochs=EPOCHS,
optimizer=SPSAOptimizer,
optim_hyperparams={'a': 0.05, 'c': 0.06, 'A':0.01*EPOCHS},
evaluate_functions={'acc': acc},
evaluate_on_train=True,
verbose = 'text',
seed= SEED
)
trainer.fit(train_dataset, val_dataset, logging_step=1)
test_acc = acc(model(test_circuits), test_labels)
print('Test accuracy:', test_acc)
##############################################################################
# import matplotlib.pyplot as plt
# fig, ((ax_tl, ax_tr), (ax_bl, ax_br)) = plt.subplots(2, 2, sharex=True, sharey='row', figsize=(10, 6))
# ax_tl.set_title('Training set')
# ax_tr.set_title('Development set')
# ax_bl.set_xlabel('Iterations')
# ax_br.set_xlabel('Iterations')
# ax_bl.set_ylabel('Accuracy')
# ax_tl.set_ylabel('Loss')
# colours = iter(plt.rcParams['axes.prop_cycle'].by_key()['color'])
# ax_tl.plot(trainer.train_epoch_costs, color=next(colours))
# ax_bl.plot(trainer.train_results['acc'], color=next(colours))
# ax_tr.plot(trainer.val_costs, color=next(colours))
# ax_br.plot(trainer.val_results['acc'], color=next(colours))
# test_acc = acc(model(test_circuits), test_labels)
# print('Test accuracy:', test_acc)
# # plt.show()
# plt.savefig('exp1_spider_a005_new_'+str(SEED)+'.png')