菜单
Installs Python packages and research code from GitHub on AIRE HPC. Use for pip install git+https, editable installs (-e), git submodules, private repositories, CUDA/C++ extension builds, or import failures between login and compute nodes on AIRE.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | aire-github-installs |
| description | Installs Python packages and research code from GitHub on AIRE HPC. Use for pip install git+https, editable installs (-e), git submodules, private repositories, CUDA/C++ extension builds, or import failures between login and compute nodes on AIRE. |
Optimize for correct node type, pinned commits, and CUDA build hygiene.
Execution note: task type (install, editable, submodule, private, cuda-build, debug), public vs private repo, compile required?, code location ($HOME vs $SCRATCH).
Read: $AIRE/knowledge/ml-on-aire.md, $AIRE/docs/software/compilers/cuda.md, $AIRE/knowledge/troubleshooting.md, $AIRE/AGENTS.md.
srun with GPU if CUDA needed.module load cuda/12.6.2; set CPATH=$CUDA_HOME/include.pip install.environment.yaml under pip: or requirements-git.txt.cd "${SCRATCH:-/mnt/scratch/$USER}/src"
git clone --recursive https://github.com/org/repo.git
cd repo && git submodule update --init --recursive
srun --partition=gpu --gres=gpu:1 --cpus-per-task=8 --mem=32G --time=2:00:00 --pty bash
module load cuda/12.6.2
module load miniforge/24.7.1
source "$(conda info --base)/etc/profile.d/conda.sh"
conda activate <env>
export CPATH="${CUDA_HOME}/include${CPATH:+:${CPATH}}"
export MAX_JOBS="${SLURM_CPUS_PER_TASK:-4}"
pip install -e ".[dev]"
python -c "import mypackage; print(mypackage.__file__)"
- pip:
- monai @ git+https://github.com/Project-MONAI/MONAI.git@v1.3.0
| Approach | AIRE fit |
|---|---|
HTTPS + GITHUB_TOKEN env (not in git) | Best on-cluster |
Clone locally → rsync to $SCRATCH | No token on cluster |
git@github.com:... with user SSH key | Not supported |
Install once during setup; in jobs only activate:
module load cuda/12.6.2
module load miniforge/24.7.1
source "$(conda info --base)/etc/profile.d/conda.sh"
conda activate <env>
python train.py
Do not pip install in every array task unless unavoidable.
| Symptom | Fix |
|---|---|
No module named X in job | Add conda block; validate script |
| CUDA headers missing | export CPATH=$CUDA_HOME/include |
No kernel image available | PyTorch ≥2.1, pytorch-cuda=12.4, sm_89 |
| Build OK on login, fail in job | Repeat module load + activate in SBATCH |
| Empty submodule dir | git submodule update --init --recursive |
pip install git+... on login for GPU extensions$TMP_SHARED~/.bashrc conda init suffices in batchaire-conda-environmentsaire-l40s-distributed-trainingaire-agent-workflowDesigns PyTorch distributed training on AIRE NVIDIA L40S GPUs (Ada, 48GB, PCIe, max 3 GPUs per node). Covers DDP vs FSDP, bf16, batch sizing, NCCL and torchrun with Slurm, TMP_SHARED staging, and SCRATCH checkpointing. Use for multi-GPU or multi-node training on AIRE gpu partition.
Orchestrates work on the University of Leeds AIRE HPC cluster using the aire-agent toolkit. Use when starting AIRE sessions, submitting Slurm jobs, choosing between raw Slurm and aire-agent tools, wiring MCP for Claude Code, or following generate-validate-submit-debug loops on AIRE.
Manages Miniforge/conda environments on the University of Leeds AIRE HPC cluster. Use when creating, updating, exporting, or debugging Python/R conda envs on AIRE; wiring activation into SBATCH or srun; fixing HOME quota; or choosing Miniforge, conda-pack, Apptainer, or Pixi for ML stacks.
Diagnoses PyTorch DDP and NCCL failures on AIRE L40S GPU jobs. Use when distributed training hangs, NCCL errors or timeouts occur, only one GPU is active, Slurm resources do not match torchrun, or seff shows wasted GPUs on AIRE.
Scaffolds and maintains modular, reproducible ML research repositories for the AIRE HPC cluster. Use for project layout (src/, configs/, scripts/, tests/), Hydra or YAML configuration, git hygiene, SCRATCH-aware paths, and integrating runs with aire-agent log_experiment. Complements domain skills such as medical-cv-research-engineer.