Dear Kwant authors I am Yongping Du from Nanjing University of Science and Technology. I have been using the KPM method to calculate the transport properties of various graphene nanostructures, together with Prof. A. P. Jauho. As a benchmark, I calculated the conductance and density of states of the Haldane model. To my surprise, I found some differences between my numerical results and the ones obtained with Kwant. I carefully checked every term of Haldane model in my implementation and found that the hopping direction of two next nearest neighbor hopping terms was the key to those differences. By reversing the direction of the hopping term in my program, I could reproduce the results from Kwant. However, I can find no physical justification for this change. I would be most grateful if you can help me to remove this discrepancy. Best regards Yongping Du