Install tomotwin by mamba

[kimdn]

Install tomotwin by mamba

[thorstenwagner]

No, I think the 3 years outdated mamba is the problem.

Maybe you want to install a more recent version? The one in pip that you have installed is not maintained anymore.

Alternatively you could try:

mamba create -n tomotwin -c nvidia -c pytorch -c rapidsai -c conda-forge "pytorch>=2.1" torchvision "pandas<2" scipy numpy matplotlib pytables cuml=23.04 cudatoolkit=11.8 "protobuf>3.20" tensorboard optuna mysql-connector-python
conda activate tomotwin
pip install tomotwin-cryoet

[kimdn]

Thanks. Indeed mamba 0.11.2 is too old.

By
./Miniforge-pypy3-Linux-x86_64.sh
I got

(base) [kimd999@deception01 bin]$ mamba --version
mamba 1.4.2
conda 23.3.1

However,
both
mamba env create -n tomotwin -f https://raw.githubusercontent.com/MPI-Dortmund/tomotwin-cryoet/main/conda_env_tomotwin.yml
and
mamba create -n tomotwin_alternative -c nvidia -c pytorch -c rapidsai -c conda-forge "pytorch>=2.1" torchvision "pandas<2" scipy numpy matplotlib pytables cuml=23.04 cudatoolkit=11.8 "protobuf>3.20" tensorboard optuna mysql-connector-python
gave me

in all three cases

1. I did 'module purge'
2. I didn't do 'module purge' so

[kimd999@deception04 ~]$ module list
Currently Loaded Modulefiles:

1. htop/2.2.0 2) tmux/3.3a 3) htop/3.2.2 4) sbank/1.2 5) cuda/9.2.148 6) nvtop/1.0

3. I did 'module purge' and 'module load cuda/11.8'
   so

(base) [kimd999@deception01 tomotwin-cryoet]$ module list
Currently Loaded Modulefiles:

1. cuda/11.8

[kimd999@deception04 ~]$ nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2022 NVIDIA Corporation
Built on Wed_Sep_21_10:33:58_PDT_2022
Cuda compilation tools, release 11.8, V11.8.89
Build cuda_11.8.r11.8/compiler.31833905_0

Even with

(base) [kimd999@deception04 ~]$ mamba --version
mamba 1.5.3
conda 23.10.0

I got the same result.

[kimdn]

Finally, I installed TomoTwin and napari-tomotwin

and opened napari-tomotwin

The solutions were

removing most contents of ~/.condarc
from

[kimd999@tlogin1 ~]$ cat .condarc
auto_update_conda: false
auto_activate_base: false
channels:
  - defaults
  - conda-forge
  - nodefaults
channel_priority: strict

to

[kimd999@tlogin1 ~]$ cat .condarc
auto_activate_base: true

and cleaning old caches by
mamba clean --all

GPT4 taught me that I need to solve pre-existing conda channel priority issues when I reported

Also, minor cache related warnings indicated that I'd better do
mamba clean --all

After removing most contents in ~/.condarc
and running mamba clean --all,
I didn't see any warnings and errors anymore when I re-tried
mamba env create -n tomotwin -f https://raw.githubusercontent.com/MPI-Dortmund/tomotwin-cryoet/main/conda_env_tomotwin.yml