This method look for all available codebooks. Those codebooks differ only in its size. The quantization type (and also quantization vector) is equal in all found cache files.