for i in index_list:
result_t += td[i]
result_v += vd[i]
elif mode == "tensor":
tx, _, vx, _ = load_and_split_mnist_tensor()
result_t = tx[index_list[0]]
result_v = vx[index_list[0]]
index_list = index_list[1:]
for i in index_list:
torch.cat((result_t, tx[i]), dim=0)
torch.cat((result_v, vx[i]), dim=0)
else:
return None
return result_t, result_v
if __name__ == "__main__":
import time
td, vd = concat_data([1, 2, 3, 4])
learner = Learner(DataLoader(td, batch_size=64, shuffle=True, num_workers=4), DataLoader(
vd, batch_size=64, shuffle=True, num_workers=4), log_interval=100, lr=0.005)
st = time.time()
learner.learn(2)
print(f"total time: {time.time() - st}")
td, vd = concat_data([3, 4, 5, 6])
learner.test_loader = DataLoader(
vd, batch_size=64, shuffle=True, num_workers=4)
learner._test()
