first version

README.md

 # Galaxy Classification
 
-
-
 ## Getting started
-
-To make it easy for you to get started with GitLab, here's a list of recommended next steps.
-
-Already a pro? Just edit this README.md and make it your own. Want to make it easy? [Use the template at the bottom](#editing-this-readme)!
-
-## Add your files
-
-- [ ] [Create](https://docs.gitlab.com/ee/user/project/repository/web_editor.html#create-a-file) or [upload](https://docs.gitlab.com/ee/user/project/repository/web_editor.html#upload-a-file) files
-- [ ] [Add files using the command line](https://docs.gitlab.com/ee/gitlab-basics/add-file.html#add-a-file-using-the-command-line) or push an existing Git repository with the following command:
-
+Clone the repository
 ```
-cd existing_repo
-git remote add origin https://gitlab.jsc.fz-juelich.de/sdlaml/tutorials/galaxy-classification.git
-git branch -M main
-git push -uf origin main
+cd /p/project1/training2347/<username> 
+git clone https://gitlab.jsc.fz-juelich.de/sdlaml/tutorials/galaxy-classification
+cd galaxy-classification
 ```
 
-## Integrate with your tools
-
-- [ ] [Set up project integrations](https://gitlab.jsc.fz-juelich.de/sdlaml/tutorials/galaxy-classification/-/settings/integrations)
-
-## Collaborate with your team
-
-- [ ] [Invite team members and collaborators](https://docs.gitlab.com/ee/user/project/members/)
-- [ ] [Create a new merge request](https://docs.gitlab.com/ee/user/project/merge_requests/creating_merge_requests.html)
-- [ ] [Automatically close issues from merge requests](https://docs.gitlab.com/ee/user/project/issues/managing_issues.html#closing-issues-automatically)
-- [ ] [Enable merge request approvals](https://docs.gitlab.com/ee/user/project/merge_requests/approvals/)
-- [ ] [Set auto-merge](https://docs.gitlab.com/ee/user/project/merge_requests/merge_when_pipeline_succeeds.html)
-
-## Test and Deploy
-
-Use the built-in continuous integration in GitLab.
-
-- [ ] [Get started with GitLab CI/CD](https://docs.gitlab.com/ee/ci/quick_start/index.html)
-- [ ] [Analyze your code for known vulnerabilities with Static Application Security Testing (SAST)](https://docs.gitlab.com/ee/user/application_security/sast/)
-- [ ] [Deploy to Kubernetes, Amazon EC2, or Amazon ECS using Auto Deploy](https://docs.gitlab.com/ee/topics/autodevops/requirements.html)
-- [ ] [Use pull-based deployments for improved Kubernetes management](https://docs.gitlab.com/ee/user/clusters/agent/)
-- [ ] [Set up protected environments](https://docs.gitlab.com/ee/ci/environments/protected_environments.html)
-
-***
-
-# Editing this README
-
-When you're ready to make this README your own, just edit this file and use the handy template below (or feel free to structure it however you want - this is just a starting point!). Thanks to [makeareadme.com](https://www.makeareadme.com/) for this template.
-
-## Suggestions for a good README
-
-Every project is different, so consider which of these sections apply to yours. The sections used in the template are suggestions for most open source projects. Also keep in mind that while a README can be too long and detailed, too long is better than too short. If you think your README is too long, consider utilizing another form of documentation rather than cutting out information.
-
-## Name
-Choose a self-explaining name for your project.
-
-## Description
-Let people know what your project can do specifically. Provide context and add a link to any reference visitors might be unfamiliar with. A list of Features or a Background subsection can also be added here. If there are alternatives to your project, this is a good place to list differentiating factors.
-
-## Badges
-On some READMEs, you may see small images that convey metadata, such as whether or not all the tests are passing for the project. You can use Shields to add some to your README. Many services also have instructions for adding a badge.
-
-## Visuals
-Depending on what you are making, it can be a good idea to include screenshots or even a video (you'll frequently see GIFs rather than actual videos). Tools like ttygif can help, but check out Asciinema for a more sophisticated method.
-
-## Installation
-Within a particular ecosystem, there may be a common way of installing things, such as using Yarn, NuGet, or Homebrew. However, consider the possibility that whoever is reading your README is a novice and would like more guidance. Listing specific steps helps remove ambiguity and gets people to using your project as quickly as possible. If it only runs in a specific context like a particular programming language version or operating system or has dependencies that have to be installed manually, also add a Requirements subsection.
-
-## Usage
-Use examples liberally, and show the expected output if you can. It's helpful to have inline the smallest example of usage that you can demonstrate, while providing links to more sophisticated examples if they are too long to reasonably include in the README.
-
-## Support
-Tell people where they can go to for help. It can be any combination of an issue tracker, a chat room, an email address, etc.
+## Environment
+The repo uses a combination of easybuild modules and a python virtual environment. You can create your own environment by executing
+```
+./setup.sh
+```
+This will load the modules given in `modules.sh` and create a python virtual environment. To load the both the modules and the virtual environment, source the activation script
+```
+source activate.sh
+```
 
-## Roadmap
-If you have ideas for releases in the future, it is a good idea to list them in the README.
+The lazy one can just activate Stefan's script:
+```
+ source /p/project1/training2437/kesselheim1/galaxy-classification/activate.sh
+```
 
-## Contributing
-State if you are open to contributions and what your requirements are for accepting them.
+## Single GPU Training
+The training on a single GPU can best be seen interactively. You can start an interactive shell with the script `start_interactive_gpu_job.sh`. To maximize execution speed, we copy the data in the ramdisk and extract the data there with the script `copy_and_unpack_data_to_shm.sh`. 
 
-For people who want to make changes to your project, it's helpful to have some documentation on how to get started. Perhaps there is a script that they should run or some environment variables that they need to set. Make these steps explicit. These instructions could also be useful to your future self.
+Then you can source the environment and run a single gpu training by invoking `python train.py`. 
 
-You can also document commands to lint the code or run tests. These steps help to ensure high code quality and reduce the likelihood that the changes inadvertently break something. Having instructions for running tests is especially helpful if it requires external setup, such as starting a Selenium server for testing in a browser.
+In summary: 
+```
+./start_interactive_gpu_job.sh
+./copy_and_unpack_data_to_shm.sh
+python train.py
+```
 
-## Authors and acknowledgment
-Show your appreciation to those who have contributed to the project.
+## Multi-GPU Training on two nodes. 
+The training on two nodes with a total of eight GPUs is started by submitting the script `run.sbatch`
+```
+sbatch run.sbatch
+```
 
-## License
-For open source projects, say how it is licensed.
 
-## Project status
-If you have run out of energy or time for your project, put a note at the top of the README saying that development has slowed down or stopped completely. Someone may choose to fork your project or volunteer to step in as a maintainer or owner, allowing your project to keep going. You can also make an explicit request for maintainers.

activate.sh

+#!/usr/bin/env bash
+
+if [ -z "$curr_dir" ]; then
+    curr_file="${BASH_SOURCE[0]:-${(%):-%x}}"
+    curr_dir="$(dirname "$curr_file")"
+fi
+
+venv_dir="$curr_dir"/env-"$SYSTEMNAME"
+
+[ -x "$(command -v deactivate)" ] && deactivate
+
+module --force purge
+if ! [ -f "$curr_dir"/modules.sh ]; then
+    echo "Cannot find \`$curr_dir/modules.sh\`; its existence is required."
+    exit 1
+fi
+source "$curr_dir"/modules.sh
+
+if ! [ -d "$venv_dir" ]; then
+    echo "Please manually execute \`nice bash ${curr_dir@Q}/set_up.sh\`" \
+         "on a login node to create the \`venv\`."
+    exit 1
+elif ! [ -f "$venv_dir"/bin/activate ]; then
+    echo "Something seems to be wrong with the \`venv\` at \`$venv_dir\`." \
+         "Please delete it (\`nice rm -rf ${venv_dir@Q}\`) and" \
+         "execute \`nice bash ${curr_dir@Q}/set_up.sh\`" \
+         "on a login node to re-create the \`venv\`."
+    exit 1
+fi
+source "$venv_dir"/bin/activate

copy_and_unpack_data_to_shm.sh

+#!/bin/bash
+mkdir /dev/shm/data
+cp  /p/project1/training2437/kesselheim1/galaxy-classification/data/* /dev/shm/data
+cd /dev/shm/data
+for x in *.tar.gz; do tar -xf $x; done
+cd -

modules.sh

+#!/usr/bin/env sh
+
+module load Stages
+module load GCC OpenMPI PyTorch torchvision IPython

requirements.txt

+torch>=1.9,<3
+torchrun_jsc>=0.0.5
+torchvision>=0.13
+galaxy-datasets
+pytorch-lightning
+scikit-learn

run.sbatch

+#!/usr/bin/env bash
+
+#SBATCH --account=atmlaml
+#SBATCH --partition=develbooster
+#SBATCH --nodes=1
+#SBATCH --ntasks-per-node=1
+#SBATCH --cpus-per-task=48
+# Use only physical cores. (Can use up to 2 threads per core.)
+#SBATCH --threads-per-core=1
+#SBATCH --gres=gpu:4
+#SBATCH --time=00:15:00
+
+curr_file="$(scontrol show job "$SLURM_JOB_ID" | grep '^[[:space:]]*Command=' | head -n 1 | cut -d '=' -f 2-)"
+curr_dir="$(dirname "$curr_file")"
+
+# Propagate the specified number of CPUs per task to each `srun`.
+export SRUN_CPUS_PER_TASK="$SLURM_CPUS_PER_TASK"
+
+source "$curr_dir"/activate.sh
+
+export MASTER_ADDR="$(scontrol show hostnames "$SLURM_JOB_NODELIST" | head -n 1)"
+if [ "$SYSTEMNAME" = juwelsbooster ] \
+       || [ "$SYSTEMNAME" = juwels ] \
+       || [ "$SYSTEMNAME" = jurecadc ] \
+       || [ "$SYSTEMNAME" = jusuf ]; then
+    # Allow communication over InfiniBand cells on JSC machines.
+    MASTER_ADDR="$MASTER_ADDR"i
+fi
+export MASTER_PORT=54123
+
+# Prevent NCCL not figuring out how to initialize.
+export NCCL_SOCKET_IFNAME=ib0
+# Prevent GLOO not being able to communicate.
+export GLOO_SOCKET_IFNAME=ib0
+
+srun env -u CUDA_VISIBLE_DEVICES python -u -m torchrun_jsc \
+       --nproc_per_node=gpu \
+       --nnodes="$SLURM_JOB_NUM_NODES" \
+       --rdzv_id="$SLURM_JOB_ID" \
+       --rdzv_endpoint="$MASTER_ADDR":"$MASTER_PORT" \
+       --rdzv_backend=c10d \
+       "$curr_dir"/main.py "$@"

setup.sh

+#!/usr/bin/env bash
+
+curr_file="${BASH_SOURCE[0]:-${(%):-%x}}"
+curr_dir="$(dirname "$curr_file")"
+
+venv_dir="$curr_dir"/env-"$SYSTEMNAME"
+
+if ! [ -d "$venv_dir" ]; then
+    [ -x "$(command -v deactivate)" ] && deactivate
+
+    module --force purge
+    if ! [ -f "$curr_dir"/modules.sh ]; then
+        echo "Cannot find \`$curr_dir/modules.sh\`; its existence is required."
+        exit 1
+    fi
+    source "$curr_dir"/modules.sh
+
+    python3 -m venv --system-site-packages "$venv_dir"
+
+    if ! [ -f "$venv_dir"/bin/activate ]; then
+        echo "Something seems to be wrong with the \`venv\` at \`$venv_dir\`." \
+             "Please delete it (\`nice rm -rf ${venv_dir@Q}\`) and" \
+             "execute \`nice bash ${curr_dir@Q}/set_up.sh\`" \
+             "on a login node to re-create the \`venv\`."
+        exit 1
+    fi
+    source "$venv_dir"/bin/activate
+
+    python -m pip install -U pip
+    if ! [ -f "$curr_dir"/requirements.txt ]; then
+        echo "Cannot find \`$curr_dir/requirements.txt\`;" \
+             "its existence is required."
+        exit 1
+    fi
+    python -m pip install -r "$curr_dir"/requirements.txt
+
+    deactivate
+else
+    echo "\`venv\` is already set up at \`$venv_dir\`. Please" \
+         "delete it (\`nice rm -rf ${venv_dir@Q}\`) to force a re-generation."
+fi

start_interactive_gpu_job.sh

+#!/bin/bash
+srun --pty --nodes=1 -A training2437 --cpu-bind=none --partition dc-gpu-devel --gres gpu:4 --time=01:10:00  /bin/bash 

train.py

+import galaxy_datasets 
+from PIL import Image
+import torch
+import torchvision
+import torchvision.models
+import types
+import os
+
+root='/dev/shm/data/'
+modelsroot='/p/project1/training2437/kesselheim1/galaxy-classification/models/'
+os.environ['TORCH_HOME']=modelsroot
+
+catalog, label_cols = galaxy_datasets.gz2(
+     root=root,
+     train=True,
+     download=True
+)
+
+from galaxy_datasets.pytorch.galaxy_datamodule import GalaxyDataModule
+
+datamodule = GalaxyDataModule(
+    label_cols=['spiral-winding-gz2_tight'],
+    catalog=catalog,
+    batch_size=256,
+    num_workers=16
+    # optional args to specify augmentations
+)
+
+datamodule.prepare_data()
+datamodule.setup()
+
+for images, labels in datamodule.train_dataloader():
+    print(images.shape, labels.shape)
+    break
+
+def train(args, model, device, train_loader, optimizer, epoch, loss_fn):
+    model.train()
+    for batch_idx, (data, target) in enumerate(train_loader):
+        target=(target>0.).type(torch.long)
+
+        data, target = data.to(device), target.to(device)
+        optimizer.zero_grad()
+        output = model(data)
+        loss = loss_fn(output, target)
+        loss.backward()
+        optimizer.step()
+        if batch_idx % args.log_interval == 0:
+            print('Train Epoch: {} [{}/{} ({:.0f}%)]\tLoss: {:.6f}'.format(
+                epoch, batch_idx * len(data), len(train_loader.dataset),
+                100. * batch_idx / len(train_loader), loss.item()))
+            if args.dry_run:
+                break
+
+def test(model, device, test_loader, loss_fn):
+    model.eval()
+    test_loss = 0
+    correct = 0
+    with torch.no_grad():
+        for data, target in test_loader:
+            target=(target>0.).type(torch.long)
+
+            data, target = data.to(device), target.to(device)
+            output = model(data)
+            test_loss += loss_fn(output, target).item()  # sum up batch loss
+            pred = output.argmax(dim=1, keepdim=True)  # get the index of the max log-probability
+            correct += pred.eq(target.view_as(pred)).sum().item()
+
+    test_loss /= len(test_loader.dataset)
+
+    print('\nTest set: Average loss: {:.4f}, Accuracy: {}/{} ({:.0f}%)\n'.format(
+        test_loss, correct, len(test_loader.dataset),
+        100. * correct / len(test_loader.dataset)))
+
+
+num_classes=2
+
+model=torchvision.models.resnet18(weights=torchvision.models.ResNet18_Weights.IMAGENET1K_V1)
+model.fc = torch.nn.Linear(model.fc.in_features, num_classes)
+model=model.cuda()
+
+loss_fn = torch.nn.CrossEntropyLoss()
+
+optimizer=torch.optim.Adam(model.parameters(), lr=0.01)
+
+num_epochs=20
+args=types.SimpleNamespace(dry_run=False, log_interval=16)
+device="cuda"
+for epoch in range(num_epochs):
+    train(args, model, device, datamodule.train_dataloader(), optimizer, epoch, loss_fn)
+    test(model, device, datamodule.val_dataloader(), loss_fn)
+