Buckets:

glennmatlin's picture
download
raw
1.02 kB
#!/usr/bin/env bash
#SBATCH --job-name=extract_ids
#SBATCH --array=0-497%15
#SBATCH --cpus-per-task=4
#SBATCH --mem=16G
#SBATCH --time=04:00:00
#SBATCH -C localNVMe
#SBATCH --output=logs/dedup/%x_%A_%a.out
#SBATCH --error=logs/dedup/%x_%A_%a.err
set -euo pipefail
REPO_DIR="${SLURM_SUBMIT_DIR:-$(cd "$(dirname "$0")/../../.." && pwd)}"
cd "$REPO_DIR"
mkdir -p logs/dedup
if [ -f .venv/bin/activate ]; then
source .venv/bin/activate
fi
export PYTHONPATH="${REPO_DIR}/src${PYTHONPATH:+:$PYTHONPATH}"
export HF_TOKEN="${HF_TOKEN:-$(cat ~/.hf_token 2>/dev/null || true)}"
python3 scripts/slurm/dedup/extract_ids.py \
--dataset "${DATASET_ID:-allenai/dolma3_mix-6T-1025-7B}" \
--manifest "${SHARD_MANIFEST:-scripts/slurm/shard_manifest_6t.txt}" \
--output-dir "${IDS_OUT_DIR:-/storage/ice-shared/cs7634/staff/TDA/soc-90/ids}" \
--task-id "${SLURM_ARRAY_TASK_ID}" \
--task-count "${TASK_COUNT:-498}" \
--shards-per-task "${SHARDS_PER_TASK:-132}" \
--startup-delay "${STARTUP_DELAY:-120}"

Xet Storage Details

Size:
1.02 kB
·
Xet hash:
57a00488a576f60e47ed3cece4f8e646b4d38353626a048777b2cb4aa5ff68e7

Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.