Upload code/scripts

42c0d23 verified 6 months ago

2.24 kB

#!/usr/bin/env bash
# scripts/launch.sh — launch a single training run, detached via nohup.
#
# Usage:
#   bash scripts/launch.sh [condition] [n_train] [seed] [gpu]
#
# Defaults:
#   condition = grokking
#   n_train   = 500
#   seed      = 42
#   gpu       = 0
#
# Survives SSH disconnects. All artifacts in experiments/runs/<run_id>/.

set -euo pipefail

ROOT="$(cd "$(dirname "$0")/.." && pwd)"
cd "${ROOT}"
source "${ROOT}/scripts/lib/nohup_runner.sh"
source "${ROOT}/scripts/lib/env.sh"
resolve_python

CONDITION="${1:-grokking}"
N_TRAIN="${2:-500}"
SEED="${3:-42}"
GPU="${4:-0}"
WANDB_MODE_ARG="${WANDB_MODE:-online}"   # online|offline|disabled
EXTRA_ARGS="${EXTRA_ARGS:-}"             # extra trainer flags pass-through
                                         # e.g. EXTRA_ARGS="--spurious_rho 0.8 --weight_decay 5e-3"
RUN_TAG="${RUN_TAG:-}"                   # extra slug inserted into the run_id between
                                         # <condition> and n<N>. Caller-controlled because
                                         # the launcher pre-creates the run_dir before the
                                         # trainer can self-tag.
                                         # e.g. RUN_TAG=spurious08  RUN_TAG=spurious095_ps10

STAMP="$(date -u +%Y%m%d-%H%M%S)"
TAG_INFIX=""
[[ -n "${RUN_TAG}" ]] && TAG_INFIX="_${RUN_TAG}"
RUN_ID="${STAMP}_${CONDITION}${TAG_INFIX}_n${N_TRAIN}_s${SEED}"
RUN_DIR="experiments/runs/${RUN_ID}"

mkdir -p "${RUN_DIR}"/{logs,results,checkpoints,figures}

echo "Launching CausalGrok run"
echo "  condition : ${CONDITION}"
echo "  n_train   : ${N_TRAIN}"
echo "  seed      : ${SEED}"
echo "  GPU       : ${GPU}"
echo "  run_id    : ${RUN_ID}"

# Build the trainer argv. EXTRA_ARGS is intentionally unquoted so it
# word-splits into individual flags (env vars cannot carry arrays).
read -r -a _extra <<<"${EXTRA_ARGS}"

CUDA_VISIBLE_DEVICES="${GPU}" launch_detached "${RUN_DIR}" \
    "${PYTHON}" -u -m experiments.causalgrok_baseline \
        --condition  "${CONDITION}" \
        --n_train    "${N_TRAIN}" \
        --seed       "${SEED}" \
        --run_dir    "${RUN_DIR}" \
        --wandb_project causalgrok \
        --wandb_mode "${WANDB_MODE_ARG}" \
        "${_extra[@]}"

	#!/usr/bin/env bash
	# scripts/launch.sh — launch a single training run, detached via nohup.
	#
	# Usage:
	# bash scripts/launch.sh [condition] [n_train] [seed] [gpu]
	#
	# Defaults:
	# condition = grokking
	# n_train = 500
	# seed = 42
	# gpu = 0
	#
	# Survives SSH disconnects. All artifacts in experiments/runs/<run_id>/.

	set -euo pipefail

	ROOT="$(cd "$(dirname "$0")/.." && pwd)"
	cd "${ROOT}"
	source "${ROOT}/scripts/lib/nohup_runner.sh"
	source "${ROOT}/scripts/lib/env.sh"
	resolve_python

	CONDITION="${1:-grokking}"
	N_TRAIN="${2:-500}"
	SEED="${3:-42}"
	GPU="${4:-0}"
	WANDB_MODE_ARG="${WANDB_MODE:-online}" # online\|offline\|disabled
	EXTRA_ARGS="${EXTRA_ARGS:-}" # extra trainer flags pass-through
	# e.g. EXTRA_ARGS="--spurious_rho 0.8 --weight_decay 5e-3"
	RUN_TAG="${RUN_TAG:-}" # extra slug inserted into the run_id between
	# <condition> and n<N>. Caller-controlled because
	# the launcher pre-creates the run_dir before the
	# trainer can self-tag.
	# e.g. RUN_TAG=spurious08 RUN_TAG=spurious095_ps10

	STAMP="$(date -u +%Y%m%d-%H%M%S)"
	TAG_INFIX=""
	[[ -n "${RUN_TAG}" ]] && TAG_INFIX="_${RUN_TAG}"
	RUN_ID="${STAMP}_${CONDITION}${TAG_INFIX}_n${N_TRAIN}_s${SEED}"
	RUN_DIR="experiments/runs/${RUN_ID}"

	mkdir -p "${RUN_DIR}"/{logs,results,checkpoints,figures}

	echo "Launching CausalGrok run"
	echo " condition : ${CONDITION}"
	echo " n_train : ${N_TRAIN}"
	echo " seed : ${SEED}"
	echo " GPU : ${GPU}"
	echo " run_id : ${RUN_ID}"

	# Build the trainer argv. EXTRA_ARGS is intentionally unquoted so it
	# word-splits into individual flags (env vars cannot carry arrays).
	read -r -a _extra <<<"${EXTRA_ARGS}"

	CUDA_VISIBLE_DEVICES="${GPU}" launch_detached "${RUN_DIR}" \
	"${PYTHON}" -u -m experiments.causalgrok_baseline \
	--condition "${CONDITION}" \
	--n_train "${N_TRAIN}" \
	--seed "${SEED}" \
	--run_dir "${RUN_DIR}" \
	--wandb_project causalgrok \
	--wandb_mode "${WANDB_MODE_ARG}" \
	"${_extra[@]}"