The ARI result of DLPFC 151676 dataset

[HappyLife-together]

Following the tutorial you provided, I ran your method on DLPFC151676 dataset, and the ARI obtained from my training was only 0.58. Deviations from the ARI results（ARI=0.60） in your tutorial （https://stagate.readthedocs.io/en/latest/T1_DLPFC.html）. And the visualization results of PAGA also have some biases. I would like to ask what might be causing this. The data parameters are all provided as per your tutorial. The data source is downloaded from the URL provided in your tutorial. In terms of environment configuration, tensorflow==1.15.0.

[xiao-kong-long]

hello, I wanna know the version of your python and rpy2, I meet some struggles in setting env

[dknqifei]

Limited by the framework of tensorflow1, STAGATE itself has a certain randomness. We discussed the issue of randomness in our paper, which we do not think has a particularly significant effect on the results.

[HappyLife-together]

Limited by the framework of tensorflow1, STAGATE itself has a certain randomness. We discussed the issue of randomness in our paper, which we do not think has a particularly significant effect on the results.

I tried training multiple times, but got the same result, ARI=0.58. This doesn't seem like the effect of randomness.

[HappyLife-together]

Limited by the framework of tensorflow1, STAGATE itself has a certain randomness. We discussed the issue of randomness in our paper, which we do not think has a particularly significant effect on the results.

hello, I wanna know the version of your python and rpy2, I meet some struggles in setting env

In my env, python==3.7, rpy2==3.4.5, hope it helps you.

[xiao-kong-long]

thanks very much!

[ly15296191253]

It's necessary to copy the offical evn completely？I install the 'scanpy==1.9.1' and 'tensorflow==1.15.0' ,and then run the code as tutorial on the dataset DLPFC151676. Finally,I get the ARI only 0.43. I don't have a idea where the problem is,may you give me some suggestion?

[yuxiaokang-source]

I also meet the same problem with STAGATE, I also install the 'scanpy==1.9.1' and 'tensorflow==1.15.0' ,when I set 10 different random seed for STAGATE initization with 151676 dataset , the result of ARI is very unstable,

[BinDuan]

I have the similar problem. I followed the same code the tutorial gave, but I can't obtain the same ARI, too much lower than the tutorial gave.

[CNSLyk7]

I also have the similar issue, the ARI that I calculated is much lower than that in the paper. Any suggestions to solve this would be greatly appreciated!