installation issue

[DmitrySemchonok]

Dear colleagues,

I have installation issue on both centos and mac systems.

After installation, on the step 3 (conda activate model_angelo).

When I run model_angelo build -h -- nothing happens

-on centos [user@dataanalysisserver1 model_angelo]$ conda activate model_angelo (model_angelo) model_angelo build -h bash: model_angelo: command not found...

- on mac import torch ModuleNotFoundError: No module named 'torch'

Any ideas?

Thank you.

Sincerely, Dmitry

[jamaliki]

Hi @DmitrySemchonok , did you get any errors while running the installation script on centos?

Also, the problem for MacOS is that the CUDA toolkit is not available for Mac. If you want to still give this a go, you'd have to run conda install pytorch torchvision torchaudio -c pytorch in the model_angelo environment, but then you can't use GPUs, so everything will be very slow. Not recommended.

[DmitrySemchonok]

conda install pytorch torchvision torchaudio -c pytorch

Hello Kiarash @jamaliki,

I tried to install it again and then got the following error:

[user@dataanalysisserver1 ~]$ cd Documents/ [user@dataanalysisserver1 Documents]$ cd model-angelo-main/ [user@dataanalysisserver1 model-angelo-main]$ cd model_angelo/ [user@dataanalysisserver1 model_angelo]$ bash install_script.sh bash: install_script.sh: No such file or directory [user@dataanalysisserver1 model_angelo]$ cd .. [user@dataanalysisserver1 model-angelo-main]$ bash install_script.sh which: no activate in (/home/user/Data/Software/cryosparc/cryosparc_master/bin:/usr/local/cuda-11.3/bin:/home/user/Data/Software/phenix-1.20.1-4487/build/bin:/home/user/Data/Software/ccp4-8.0/etc:/home/user/Data/Software/ccp4-8.0/bin:/home/user/Data/Software/cryosparc/cryosparc_master/bin:/usr/local/cuda-11.3/bin:/home/user/Data/Software/miniconda/condabin:/sbin:/bin:/usr/bin:/usr/local/bin:/usr/local/sbin:/usr/sbin:/home/user/Data/Software/Relion/bin:/usr/lib64/openmpi/bin:/home/user/.local/bin:/home/user/bin:/home/user/Data/Software/Relion/bin:/usr/lib64/openmpi/bin) install_script.sh: line 31: source: model_angelo: is a directory Collecting package metadata (current_repodata.json): done Solving environment: done

All requested packages already installed.

Retrieving notices: ...working... done install_script.sh: line 37: pip: command not found File "setup.py", line 12 sys.path.insert(0, f"{os.path.dirname(file)}/model_angelo") ^ SyntaxError: invalid syntax Did not download weights because the flag -w or --download-weights was not specified [user@dataanalysisserver1 model-angelo-main]$

==========================================

By the way, the step 3, if I am not mistaking, should be "cd model-angelo-main" unstead of "cd model-angelo" For me the cd model-angelo bash install_script.sh

did not work.

Kind regards, Dmitry

[jamaliki]

My bad, it seems like miniconda does not have the activate binary, let me change the installation script and you can try again.

[jamaliki]

Could you pull the newest commit and try now @DmitrySemchonok

[DmitrySemchonok]

[user@dataanalysisserver1 model-angelo-main]$ bash install_script.sh which: no activate in (/home/user/Data/Software/cryosparc/cryosparc_master/bin:/usr/local/cuda-11.3/bin:/home/user/Data/Software/phenix-1.20.1-4487/build/bin:/home/user/Data/Software/ccp4-8.0/etc:/home/user/Data/Software/ccp4-8.0/bin:/home/user/Data/Software/cryosparc/cryosparc_master/bin:/usr/local/cuda-11.3/bin:/home/user/Data/Software/miniconda/condabin:/sbin:/bin:/usr/bin:/usr/local/bin:/usr/local/sbin:/usr/sbin:/home/user/Data/Software/Relion/bin:/usr/lib64/openmpi/bin:/home/user/.local/bin:/home/user/bin:/home/user/Data/Software/Relion/bin:/usr/lib64/openmpi/bin) install_script.sh: line 31: source: model_angelo: is a directory Collecting package metadata (current_repodata.json): done

Hello Kiarash @jamaliki,

1. the pulling for me doesn't work for me - I get the error (not sure why)

   [user@dataanalysisserver1 ~]$ git clone git@github.com:3dem/model-angelo.git Cloning into 'model-angelo'... Permission denied (publickey). fatal: Could not read from remote repository.

So I just downloaded it from this website.

2. step 3 has yet a typo, I suppose. After I run again the

cd model-angelo bash install_script.sh

In the main folder "model-angelo-main" where the install_script.sh I get the following error - please see below

[user@dataanalysisserver1 model_angelo]$ cd .. [user@dataanalysisserver1 model-angelo-main]$ bash install_script.sh which: no activate in (/home/user/Data/Software/cryosparc/cryosparc_master/bin:/usr/local/cuda-11.3/bin:/home/user/Data/Software/phenix-1.20.1-4487/build/bin:/home/user/Data/Software/ccp4-8.0/etc:/home/user/Data/Software/ccp4-8.0/bin:/home/user/Data/Software/cryosparc/cryosparc_master/bin:/usr/local/cuda-11.3/bin:/home/user/Data/Software/miniconda/condabin:/sbin:/bin:/usr/bin:/usr/local/bin:/usr/local/sbin:/usr/sbin:/home/user/Data/Software/Relion/bin:/usr/lib64/openmpi/bin:/home/user/.local/bin:/home/user/bin:/home/user/Data/Software/Relion/bin:/usr/lib64/openmpi/bin) install_script.sh: line 31: source: model_angelo: is a directory Collecting package metadata (current_repodata.json): done Solving environment: done

All requested packages already installed.

Retrieving notices: ...working... done install_script.sh: line 37: pip: command not found File "setup.py", line 12 sys.path.insert(0, f"{os.path.dirname(file)}/model_angelo") ^ SyntaxError: invalid syntax Did not download weights because the flag -w or --download-weights was not specified [user@dataanalysisserver1 model-angelo-main]$ conda activate model_angelo (model_angelo) model_angelo build -h bash: model_angelo: command not found... (model_angelo)

Any idea?

Thank you!

Sincerely, Dmitry

[jamaliki]

Hi, wow this is very strange. Can you delete the model-angelo that you have (the folder with the code). Change

git clone git@github.com:3dem/model-angelo.git

with

git clone https://github.com/3dem/model-angelo.git

Now, you should have a model-angelo directory. Can you try the steps again. Sorry for the issues, some of the installation steps need a few users to polish things out.

Thank you so much for being an early adopter :)

[DmitrySemchonok]

Hi, wow this is very strange. Can you delete the model-angelo that you have (the folder with the code). Change

git clone git@github.com:3dem/model-angelo.git

with

git clone https://github.com/3dem/model-angelo.git

Now, you should have a model-angelo directory. Can you try the steps again. Sorry for the issues, some of the installation steps need a few users to polish things out.

Thank you so much for being an early adopter :)

Hi, Sure no problem.

The git worked. But now we have a new error. Please see below

[jamaliki]

Sorry for all the back and forth. This is a tiny issue with conda, but I think I know what is going wrong. What happens when you run

export PATH="/home/user/Data/Software/miniconda/bin:$PATH"
bash install_script.sh

[DmitrySemchonok]

Sorry for all the back and forth. This is a tiny issue with conda, but I think I know what is going wrong. What happens when you run

export PATH="/home/user/Data/Software/miniconda/bin:$PATH"
bash install_script.sh

All good - we are progressing :)

[DmitrySemchonok]

Sorry for all the back and forth. This is a tiny issue with conda, but I think I know what is going wrong. What happens when you run

export PATH="/home/user/Data/Software/miniconda/bin:$PATH"
bash install_script.sh

All good - we are progressing :)

@jamaliki Sorry I guess I have to be in file mode_angelo. So now it is compiling. I will update you about the results

[DmitrySemchonok]

@jamaliki

Hi, It seems to work!

Thank you!

Sincerely, Dmitry

[jamaliki]

Great, thanks for the patience @DmitrySemchonok ! Let me know if you have any other problems :)