svm_cv.py won't work with tanimoto kernel

[LanceKnight]

Dear professor,

This is just a reference for me in the future and also for people who encounter the same problem as I did.

It seems like the codes won't work with Tanimoto kernel. It was caused by the list type nature of input data. Adding a conditional statement to convert inputs from list to ndarray for both the training and testing should work:

For training:

		if kwargs['kernel'] not in ['tanimoto']:
			model.fit([x[0] for x in data[0]['mols']], data[0]['y'])
		else:
			train_x = np.array([x[0] for x in data[0]['mols']])
			train_y = np.array(data[0]['y'])

For testing:

	if kwargs['kernel'] not in ['tanimoto']:		
		predicted_y = model.predict(test_x)
	else:
		test_x = np.array(test_x)
		predicted_y = model.predict(test_x)