Failed to preprocess about deeph-pack HOT 9 CLOSED

Hi,

Could you provide the output of deeph-preprocess? This should include the specific path to the error.log file, which contains information about the error.

Alternatively, you could try running the following command to see the contents of all error.log files:

cat /public/wcd/twisted/example/work_dir/dataset/processed/*/error.log

Once you've located the error.log file, please copy and paste its contents here so we can know what went wrong.

from deeph-pack.

I tried versions 1.5.4 and 1.6.6, but they did not run successfully. I checked the file location of 1.5.4 correctly and used the pwd instruction to check the location. What should I do to solve this error?
error(1.5.4) is :
error（1.5.4）.log
[stdout of cmd "/public/apps/julia-1.5.4/bin/juila /public/apps/miniconda3/lib/python3.9/site-packages/deeph/preprocess/openmx_get_data.jl --input_dir /public/wcd/twisted/example/work_dir/dataset/raw/0 --output_dir /public/wcd/twisted/example/work_dir/dataset/processed/0 --save_overlap false"]:
[stderr of cmd "/public/apps/julia-1.5.4/bin/juila /public/apps/miniconda3/lib/python3.9/site-packages/deeph/preprocess/openmx_get_data.jl --input_dir /public/wcd/twisted/example/work_dir/dataset/raw/0 --output_dir /public/wcd/twisted/example/work_dir/dataset/processed/0 --save_overlap false"]:
/bin/sh: /public/apps/julia-1.5.4/bin/juila: 没有那个文件或目录
error(1.6.6)is:
[stdout of cmd "/public/apps/julia-1.5.4/bin/julia /public/apps/miniconda3/lib/python3.9/site-packages/deeph/preprocess/openmx_get_data.jl --input_dir /public/wcd/twisted/example/work_dir/dataset/raw/0 --output_dir /public/wcd/twisted/example/work_dir/dataset/processed/0 --save_overlap false"]:
[stderr of cmd "/public/apps/julia-1.5.4/bin/julia /public/apps/miniconda3/lib/python3.9/site-packages/deeph/preprocess/openmx_get_data.jl --input_dir /public/wcd/twisted/example/work_dir/dataset/raw/0 --output_dir /public/wcd/twisted/example/work_dir/dataset/processed/0 --save_overlap false"]:
ERROR: LoadError: ArgumentError: Package StaticArrays not found in current path:

• Run import Pkg; Pkg.add("StaticArrays") to install the StaticArrays package.
  Stacktrace:
  [1] require(::Module, ::Symbol) at ./loading.jl:893
  [2] include(::Function, ::Module, ::String) at ./Base.jl:380
  [3] include(::Module, ::String) at ./Base.jl:368
  [4] exec_options(::Base.JLOptions) at ./client.jl:296
  [5] _start() at ./client.jl:506
  in expression starting at /public/apps/miniconda3/lib/python3.9/site-packages/deeph/preprocess/openmx_get_data.jl:1
  error（1.5.4）.log
  error(1.6.6).log

from deeph-pack.

It appears that you do not have the required Julia package installed. Run

/public/apps/julia-1.5.4/bin/juila

and then execute the command:

import Pkg
Pkg.add("StaticArrays")
Pkg.add("Arpack")
Pkg.add("HDF5")
Pkg.add("ArgParse")
Pkg.add("JLD")
Pkg.add("JSON")
Pkg.add("IterativeSolvers")
Pkg.add("DelimitedFiles")
Pkg.add("LinearMaps")

See details in https://github.com/mzjb/DeepH-pack#julia

from deeph-pack.

Thank you very much for your answer. After installing the relevant environment you mentioned, the program seems to be running normally. But I found that the environment you mentioned is missing Pardiso.jl. After I simply ran add Pardiso.jl, the program can run normally, but I did not follow the installation instructions. May I ask what impact this will have?

from deeph-pack.

@wangchengDon

Pardiso.jl is used for solving sparse linear equations, and is utilized in DeepH-pack for computing a few number of eigenvalues of large sparse matrices. Therefore, it is not required for deeph-preprocess. Only when running deeph-inference with dense_calc set to False, which involves computing a small number of eigensolutions for a large sparse matrix, will Pardiso.jl be used.

from deeph-pack.

Thank you!
I am very grateful for your kind answer

from deeph-pack.

Hello dear developer, I have a question in train. How should I set parameters to support parallel operations after running? I am setting device=CPU num_ After threads=48, only one core was used in the calculation, rather than one node and forty-eight cores

from deeph-pack.

@wangchengDon

The parameter num_threads only affects the openmp threads used by PyTorch, and increasing this value may not lead to a significant improvement in efficiency. Therefore, I suggest not setting the thread count too high.

The parallel training and inference for DeepH are still in development and will be released soon.

from deeph-pack.

Thank you very much for your patient and meticulous answer

from deeph-pack.