Skip to content
cccskills
FreePublished on GitHub

nemo-rl-brev-etiquette

Organize NeMo-RL experiments on Brev by placing large outputs and shared caches on /ephemeral, loading credentials without exposing them, checking disk capacity, and keeping concise reproduction records in the source checkout.

OriginalShow the original description

Brev instance operating guidance for NeMo-RL agents working in /home/ubuntu/RL with limited workspace disk, a larger /ephemeral volume, and optional /home/ubuntu/RL/.env secrets. Use when running nemo-rl-auto-research campaigns, experiments, training jobs, model or dataset downloads, shared cache-heavy commands, log-producing runs, checkpoint generation, W&B or Hugging Face authenticated workflows, or any workflow that may create large files on Brev.

Works withHugging Face
See how to install

Good for

  • Prepare storage before a long training run
  • Share model and dataset caches across experiments
  • Run authenticated W&B or Hugging Face workflows
  • Clean experiment outputs while retaining reproduction records

Files included (5)

  • SKILL.md5.1 KB
  • BENCHMARK.md4.2 KB
  • evals/evals.json2.2 KB
  • skill-card.md3.8 KB
  • skill.oms.sig4.5 KB

SKILL.md (original)

Review the instructions the agent will receive before you install.

Brev Etiquette

Operate as though /home/ubuntu/RL is the source checkout and /ephemeral is the working storage for generated experiment state. Keep the repo small, reproducible, and easy to inspect. Move bulky run outputs to /ephemeral before launching anything expensive.

Storage Rules

  • Keep code edits, small config changes, committed experiment hypotheses, and concise reproducibility records under /home/ubuntu/RL.
  • Put generated experiment assets under /ephemeral, including checkpoints, run logs, Ray temp directories, W&B offline files, profiler traces, evaluation dumps, rollout samples, and per-experiment artifacts.
  • Keep reusable caches under one shared /ephemeral cache root per user, not under each experiment. This includes Hugging Face models, dataset caches, PyTorch caches, Triton caches, uv caches, and pip caches.
  • Before a campaign or long run, check capacity with df -h /home/ubuntu/RL /ephemeral and avoid starting if /ephemeral is missing or nearly full.
  • Create a campaign root such as /ephemeral/nemo-rl/${USER:-ubuntu}/nemo-rl-auto-research/<campaign> and use one subdirectory per experiment.
  • Do not leave large files, cache directories, or generated outputs in the git checkout. If a tool defaults to the repo, override its output/cache path before running it.

Environment Secrets

  • Treat /home/ubuntu/RL/.env as the local secret store. It may contain keys such as WANDB_API_KEY, HF_TOKEN, or HUGGING_FACE_HUB_TOKEN.
  • Before any run that may need external auth, load /home/ubuntu/RL/.env when it exists. Never print, cat, log, commit, or summarize secret values.
  • If /home/ubuntu/RL/.env is absent, or a required key is still unset after loading it, remind the user to add the needed key to that file before launching authenticated work.
if [ -f /home/ubuntu/RL/.env ]; then
  set -a
  . /home/ubuntu/RL/.env
  set +a
else
  echo "Missing /home/ubuntu/RL/.env; add required keys such as WANDB_API_KEY or HF_TOKEN before authenticated runs."
fi

Auto-Research Pattern

When using nemo-rl-auto-research, keep the git ledger in the repo and heavy evidence on /ephemeral.

if [ -f /home/ubuntu/RL/.env ]; then
  set -a
  . /home/ubuntu/RL/.env
  set +a
fi

BREV_ROOT=/ephemeral/nemo-rl/${USER:-ubuntu}
CACHE_ROOT=$BREV_ROOT/cache
CAMPAIGN_ROOT=$BREV_ROOT/nemo-rl-auto-research/<campaign>
EXP_DIR=$CAMPAIGN_ROOT/<experiment>
mkdir -p "$EXP_DIR"/{logs,checkpoints,artifacts,ray,tmp,wandb}
mkdir -p "$CACHE_ROOT"/{huggingface,torch,triton,uv,pip,xdg,wandb}

export HF_HOME=$CACHE_ROOT/huggingface
export HF_HUB_CACHE=$HF_HOME/hub
export HF_DATASETS_CACHE=$HF_HOME/datasets
export TRANSFORMERS_CACHE=$HF_HOME/transformers
export TORCH_HOME=$CACHE_ROOT/torch
export TRITON_CACHE_DIR=$CACHE_ROOT/triton
export UV_CACHE_DIR=$CACHE_ROOT/uv
export PIP_CACHE_DIR=$CACHE_ROOT/pip
export XDG_CACHE_HOME=$CACHE_ROOT/xdg
export WANDB_CACHE_DIR=$CACHE_ROOT/wandb
export RAY_TMPDIR=$EXP_DIR/ray
export TMPDIR=$EXP_DIR/tmp
export WANDB_DIR=$EXP_DIR/wandb

Record the absolute /ephemeral paths in the nemo-rl-auto-research TSV fields for log path, checkpoint path, artifacts, shared cache root, and command. If the TSV itself may grow large, store the full TSV in /ephemeral and keep a small pointer file or summary in the repo.

Launch Checklist

  • Inspect disk first: df -h /home/ubuntu/RL /ephemeral.
  • Choose a unique /ephemeral run root before editing recipes or launching jobs.
  • Reuse a shared cache root such as /ephemeral/nemo-rl/${USER:-ubuntu}/cache across experiments unless a run explicitly requires a clean cache.
  • Override recipe output paths, logger paths, checkpoint paths, and temp paths to point under the experiment directory.
  • Override cache paths to point under the shared cache root.
  • Stream stdout/stderr to $EXP_DIR/logs/run.log or an equivalent file under /ephemeral.
  • Periodically check disk during long runs with df -h /ephemeral and stop gracefully if the volume is approaching exhaustion.
  • At the end, summarize the important metrics and paths in the repo ledger; do not copy bulky artifacts back into /home/ubuntu/RL.

Cleanup

  • Clean only files that belong to the current campaign or experiment.
  • Prefer pruning clearly named experiment directories under /ephemeral/nemo-rl/...; never remove shared caches or another user's run directory without an explicit instruction.
  • Preserve enough small metadata in the repo to reproduce a result after /ephemeral is cleaned.

Reviews

No reviews yet. Tell others what you thought.

Comments

No login needed. Share how it worked for you, ask a question, or note where it did not run.

No comments yet. Be the first to write one.

Comments are published right away. Do not include personal information such as email addresses or phone numbers. Up to 2 URLs.

0 / 1,000 characters

This form is protected by reCAPTCHA, and the Google Privacy Policy and Terms of Service apply.

More skills from this repository

Move pandas-style table processing to NVIDIA GPUs using cuDF. Choose minimal-change acceleration or explicit GPU code, handle joins and aggregations, compare results with pandas, and use dask-cuDF when data exceeds GPU memory.

  • Accelerate existing pandas processing with minimal changes
  • Move joins and aggregations to GPU code
  • Compare missing values and ordering with pandas
NVIDIA/skills3.7KUpdated Oct 22, 2026

Customize NVIDIA Nemotron Voice Agent’s Generic Pipecat example for appointment scheduling, five-field patient intake, or a custom healthcare conversation. Build prompts and tool handlers without a separate backend, then validate conversation behavior, services, and voice interaction.

  • Build an appointment-scheduling voice agent
  • Collect five patient-intake fields through conversation
  • Design a custom healthcare conversation with tool calls
NVIDIA/skills3.7KUpdated Oct 22, 2026

Calibrate live cameras by recording their RTSP video streams through VIOS and processing the clips with AutoMagicCalib. The workflow imports settings and alignment files, checks readiness, tracks calibration progress, and retrieves results.

  • Calibrate cameras from live RTSP streams
  • Import settings and alignment files for a camera scene
  • Retrieve evaluation statistics using ground truth data
NVIDIA/skills3.7KUpdated Oct 22, 2026

Test a running AMC service with the bundled four-camera warehouse dataset. Run camera calibration from upload to completion, then check distance and image projection errors against ground-truth data to verify the installation.

  • Verify a newly installed AMC service
  • Check the sample's calibration errors against ground truth
  • Walk through the sample workflow in Swagger UI
NVIDIA/skills3.7KUpdated Oct 22, 2026

Calibrate multiple cameras from recorded MP4 files through the AutoMagicCalib REST API. The skill handles project creation, uploads, input verification, progress checks, and evaluation when reference data is supplied, with optional VGGT refinement.

  • Calibrating cameras from synchronized MP4 recordings
  • Uploading local settings and alignment files
  • Evaluating calibration against ground-truth data
NVIDIA/skills3.7KUpdated Oct 22, 2026

Set up AutoMagicCalib from NGC release containers using Docker Compose. The skill handles registry authentication, port and path configuration, optional VGGT model download, service startup, and readiness checks for the backend and web interface.

  • Launching the AutoMagicCalib backend and web interface
  • Configuring service ports and network access
  • Adding the optional VGGT model during setup
NVIDIA/skills3.7KUpdated Oct 22, 2026

See all skills by NVIDIA

Report a problem with this skill