Issue with Replicating Results

[shuisuhi]

Hello! I am currently attempting to replicate your study published in the "Multimodal Fusion Transformer for Remote Sensing Image Classification". First, I would like to express my gratitude for your contribution to this field and for making your code available publicly. While replicating your work, I strictly followed the methodology described in your paper and used the open-source code. However, the results I obtained are somewhat different from those reported in your paper. Could you provide some details about the experimental setup, such as the version of Python and pytorch? Are there any hyperparameters that I might need to adjust to align with your experimental settings? In your opinion, what other factors might contribute to the discrepancy in replication results? Thank you once again for your work, and I look forward to your response.