search

lllxxyyy-lxy / stMCDI

stMCDI: Masked Conditional Diffusion Model with Graph Neural Network for Spatial Transcriptomics Data Imputation
9 stars 0 forks source link

Does the Code Match the Paper's Description? #2

Open mQvQ opened 2 weeks ago

mQvQ commented 2 weeks ago

The paper states that a Graph-AutoEncoder was used to obtain the latent representation. But the code appears to directly use the original gene counts after preprocessing. Could you provide code that is consistent with the description in the paper? Thanks !

fwtsinghua commented 1 week ago

SAME QUESTION!!!
IT LOOKS LIKE "only modified the code line 35 in file src/main_model_st.py " with 'self.gnnencoder = GNNModule(input_dim, input_dim)' , and never used anywhere.

The code here is strange.