search

kieranrcampbell / clonealign

Bayesian inference of clone-specific gene expression estimates by integrating single-cell RNA-seq and single-cell DNA-seq data
Apache License 2.0
32 stars 6 forks source link

module 'tensorflow' has no attribute 'reset_default_graph' #12

Open skatragadda-nygc opened 4 years ago

skatragadda-nygc commented 4 years ago

Would you know how to fix this issue?

Constructing tensorflow graph Error in py_get_attr_impl(x, name, silent) : AttributeError: module 'tensorflow' has no attribute 'reset_default_graph' Calls: run_clonealign ... py_get_attr_or_item -> py_get_attr -> py_get_attr_impl Execution halted

Thanks

kieranrcampbell commented 4 years ago

This should be fixed in 205d8b8 - can you reinstall and try again? Note that this now requires tensorflow 2.1.0, readme updated with installation instructions

Toseph commented 4 years ago

Hello, I am having a similar issue. The latest update today did resolve the "no attribute" error, but now I see this.

Error in (function (gene_expression_data, copy_number_data, max_iter = 200, : unused argument (seed = 1392) Calls: run_clonealign -> do.call Execution halted

The Rscript looks like this, `library(Matrix) library(tensorflow) library(clonealign)

Load Rdata file.

load("HPNE_clonealign.Rdata") reticulate::py_config() Sys.getenv() Sys.unsetenv('VIRTUAL_ENV') Sys.unsetenv('PATH') Sys.unsetenv('RETICULATE_REQUIRED_MODULE') Sys.getenv()

Fit model.

model_from_clonealign <- run_clonealign(ca_data$gene_expression_data, ca_data$copy_number_data, seed = 1392)

Save.`

kieranrcampbell commented 4 years ago

Hi @Toseph

I removed the seed argument to clonealign so you can now set.seed(...) beforehand as per bioconductor guidelines.

Thanks,

Kieran

Toseph commented 4 years ago

Thanks for the update @kieranrcampbell! We will test this today and get back to you.