search

qitianwu / NodeFormer

The official implementation of NeurIPS22 spotlight paper "NodeFormer: A Scalable Graph Structure Learning Transformer for Node Classification"
291 stars 28 forks source link

Same results for the last four runs #17

Closed null-xyj closed 8 months ago

null-xyj commented 10 months ago

Hello, I've encountered an issue while running the model on the official splits of Cora, CiteSeer, and Pubmed. Despite conducting five separate runs, the results for the last four runs are identical. I have meticulously ensured that the parameters are reset at the beginning of each run. Could this be a mere coincidence, or might there be an underlying error that I'm missing?

qitianwu commented 10 months ago

Does this phenomenon happen for three datasets and different machines? If not, it might be caused by a coincidence. Maybe one can also try increasing the number of runs, and see what happen for the other separate runs.