Running the SGP offline training. An error occured MemoryError: std::bad_alloc

[siinghim]

I follow the tutorial the first part of sparse_gp_tutorial.ipynb training an offline sparse gaussian process.
Once I increased the number of training_size and validation_size, an error MemoryError: std::bad_alloc will occured during the following step.
I also use my own data which contains 274 atoms and 9000 structures to train the model. And it also failed.

# Calculate descriptors of the validation and training structures.
print("Computing descriptors of validation points...")
validation_strucs = []
validation_forces = np.zeros((validation_size, noa, 3))
for n, snapshot in enumerate(validation_pts):
    pos = positions[snapshot]
    frcs = forces[snapshot]

    # Create structure object, which computes and stores descriptors.
    struc = \
        Structure(cell, coded_species, pos, cutoff, descriptors)
    validation_strucs.append(struc)
    validation_forces[n] = frcs
print("Done.")

print("Computing descriptors of training points...")
training_strucs = []
training_forces = np.zeros((training_size, noa, 3))
for n, snapshot in enumerate(training_pts):
    pos = positions[snapshot]
    frcs = forces[snapshot]

    # Create structure object, which computes and stores descriptors.
    struc = \
        Structure(cell, coded_species, pos, cutoff, descriptors)

    # Assign force labels to the training structure.
    struc.forces = frcs.reshape(-1)

    training_strucs.append(struc)
    training_forces[n] = frcs
print("Done.")

What I want to do is just using my own data (the data is large including 9000 structures and 274 atoms) to train an offline model, not on-the-fly.

I will appreciate it if there is a way to figure it out!