Something went wrong. Try again.
This repository has no description
Something went wrong. Try again.
123456789101112131415161718192021222324252627282930313233343536373839#!/usr/bin/env bash# Capacity-floor sweep: solo TRMs at decreasing hidden width, same protocol as the# reproduction run, so every point is comparable to the D=512 anchor (84.5%).## scripts/run_sweep.sh [widths...] # default: 256 128 64## Runs sequentially in one detached process; each width gets its own run name,# log, metrics and checkpoints. Progress: runs/sweep.out## This curve is the x-axis of the whole DSEM project: it says how badly a lone# unit of each size does, which is what a colony of such units must beat.set -euo pipefail
cd "$(dirname "$0")/.."
WIDTHS=("$@")[ ${#WIDTHS[@]} -eq 0 ] && WIDTHS=(256 128 64)
for W in "${WIDTHS[@]}"; do RUN="sweep-d${W}" echo "=== $(date -Is) starting ${RUN} ===" PYTHONUNBUFFERED=1 .venv/bin/python -m dsem.pretrain \ arch=trm \ data_paths="[data/sudoku-extreme-1k-aug-1000]" \ data_paths_test="[data/sudoku-testsub-12k]" \ evaluators="[]" \ epochs=50000 eval_interval=5000 \ global_batch_size=768 \ lr=1e-4 puzzle_emb_lr=1e-4 weight_decay=1.0 puzzle_emb_weight_decay=1.0 \ arch.mlp_t=True arch.pos_encodings=none \ arch.L_layers=2 arch.H_cycles=3 arch.L_cycles=6 \ "arch.hidden_size=${W}" \ ema=True \ "+run_name=${RUN}" > "runs/${RUN}.out" 2>&1 < /dev/null echo "=== $(date -Is) finished ${RUN} ==="done
echo "=== $(date -Is) sweep complete ==="