```

!/usr/bin/env bash

MiloPYP Quick tutorials

https://nextpyp.app/milopyp/quick_tutorial/

Download tutorial data in cwd and run Cellular Content Exploration

Takes ~ 40 minutes on one NVIDIA GeForce GTX 1080 Ti

James Vincent help@sbgrid.org

Nov 1, 2024

export SBGRID_ALLOW=true # sbgrid internal only

Start SBGrid environment

source /programs/sbgrid.shrc export MILOPYP_X=0.5.0_cu11.8

Set location of milopyp python scripts:

milo_dir=/programs//x86_64-linux/milopyp/${MILOPYP_X}/cet_pick/cet_pick

Download globular tutorial data | Globular-shaped particles (EMPIAR-10304)

wget https://nextpyp.app/files/data/milopyp_globular_tutorial.tbz tar xvfz milopyp_globular_tutorial.tbz mkdir data sample_data mv .txt ./data mv .rec .ali .tlt ./sample_data

-- Cellular content exploration module --

Specify GPU to use - this must be set

export CUDA_VISIBLE_DEVICES=0

Training 2d3d

python.milopyp ${milo_dir}/simsiam_main.py simsiam2d3d --num_epochs 20 --exp_id test_sample --bbox 36 --dataset simsiam2d3d --arch simsiam2d3d_18 --lr 1e-3 --train_img_txt sample_train_explore_img.txt --batch_size 256 --val_intervals 20 --save_all --gauss 0.8 --dog 3,5

Inference 2d3d

python.milopyp ${milo_dir}/simsiam_test_hm_2d3d.py simsiam2d3d --exp_id test_sample --bbox 36 --dataset simsiam2d3d --arch simsiam2d3d_18 --test_img_txt sample_train_explore_img.txt --load_model exp/simsiam2d3d/test_sample/model_20.pth --gauss 0.8 --dog 3,5