diff options
Diffstat (limited to 'scripts/run_ladder.sh')
| -rwxr-xr-x | scripts/run_ladder.sh | 4 |
1 files changed, 3 insertions, 1 deletions
diff --git a/scripts/run_ladder.sh b/scripts/run_ladder.sh index 764632a..65b43a9 100755 --- a/scripts/run_ladder.sh +++ b/scripts/run_ladder.sh @@ -10,6 +10,7 @@ # PREP_TOKENS=3e9 tokens to tokenize if DATA missing [3e9] # OUT=runs output root [runs] # MICRO_BS=8 per-GPU micro batch [8] +# SET="k=v k2=v2" extra --set overrides for every run (e.g. vocab=8192 seq_len=256) # DRY=1 print the plan and exit # # Runs are sequential (each takes the whole node), resume-safe: a finished run leaves OUT/<name>/DONE @@ -24,7 +25,8 @@ DATA=${DATA:-data/fineweb} PREP_TOKENS=${PREP_TOKENS:-3e9} OUT=${OUT:-runs} MICRO_BS=${MICRO_BS:-8} -EXTRA=${TOKENS:+--set tokens=$TOKENS} +EXTRA="--set ${TOKENS:+tokens=$TOKENS} ${SET:-}" +[ "$EXTRA" = "--set " ] && EXTRA="" echo "== zbp-scaling ladder: sizes=[$SIZES] arms=[$ARMS] nproc=$NPROC data=$DATA out=$OUT ${TOKENS:+tokens=$TOKENS}" [ "$NPROC" -ge 1 ] || { echo "no GPUs detected; set NPROC"; exit 1; } |