MOVZXandHinata
7e80cc67ee
``` [10-15 06:31:36] PP: 0.0 t/s | TG: 4.0 t/s | K/V Cache: 61.5% | Pfx Cache: 0.0% | Draft Acc: 11.1% [10-15 06:31:37] PP: 0.0 t/s | TG: 5.0 t/s | K/V Cache: 61.5% | Pfx Cache: 0.0% | Draft Acc: 50.0% [10-15 06:31:38] PP: 0.0 t/s | TG: 6.0 t/s | K/V Cache: 60.7% | Pfx Cache: 0.0% | Draft Acc: 33.3% [10-15 06:31:39] PP: 0.0 t/s | TG: 10.0 t/s | K/V Cache: 61.5% | Pfx Cache: 0.0% | Draft Acc: 77.8% [10-15 06:31:40] PP: 0.0 t/s | TG: 6.0 t/s | K/V Cache: 63.2% | Pfx Cache: 0.0% | Draft Acc: 66.7% [10-15 06:31:41] PP: 0.0 t/s | TG: 5.0 t/s | K/V Cache: 62.3% | Pfx Cache: 0.0% | Draft Acc: 22.2% [10-15 06:31:42] PP: 0.0 t/s | TG: 8.0 t/s | K/V Cache: 63.2% | Pfx Cache: 0.0% | Draft Acc: 100.0% [10-15 06:31:43] PP: 0.0 t/s | TG: 11.0 t/s | K/V Cache: 64.8% | Pfx Cache: 0.0% | Draft Acc: 88.9% ``` Co-Authored-By: Hinata <hinata@idihore.id> Signed-off-by: MOVZX <movzx@yahoo.com>
94 lines
2.8 KiB
Bash
Executable File
94 lines
2.8 KiB
Bash
Executable File
#!/usr/bin/env bash
|
|
set -uo pipefail
|
|
|
|
CONTAINER="${VLLM_CONTAINER:-vllm}"
|
|
LOG_FILE="${VLLM_LOG_FILE:-/var/log/vllm.log}"
|
|
SOURCE="docker"
|
|
TAIL=""
|
|
SINCE="10m"
|
|
|
|
usage() {
|
|
cat <<EOF
|
|
Usage: vllm-log [-f] [-c CONTAINER] [-n LINES] [-s SINCE]
|
|
-f read $LOG_FILE (tail -f) instead of docker logs
|
|
-c CONTAINER docker container name (default: $CONTAINER)
|
|
-n LINES start from last LINES (docker: --tail, file: -n). Disables -s.
|
|
-s SINCE docker logs --since (default: ${SINCE}. "" = full log)
|
|
-h help
|
|
EOF
|
|
}
|
|
|
|
while getopts "fc:n:s:h" opt; do
|
|
case "$opt" in
|
|
f) SOURCE="file" ;;
|
|
c) CONTAINER="$OPTARG" ;;
|
|
n) TAIL="$OPTARG"; SINCE="" ;;
|
|
s) SINCE="$OPTARG"; TAIL="" ;;
|
|
h) usage; exit 0 ;;
|
|
*) usage >&2; exit 2 ;;
|
|
esac
|
|
done
|
|
|
|
if [ "$SOURCE" = "file" ]; then
|
|
[ -r "$LOG_FILE" ] || { echo "cannot read $LOG_FILE" >&2; exit 1; }
|
|
if [ -n "$TAIL" ]; then
|
|
tail -n "$TAIL" -f "$LOG_FILE"
|
|
else
|
|
tail -n 200 -f "$LOG_FILE"
|
|
fi
|
|
else
|
|
if ! docker inspect --format '{{.Name}}' "$CONTAINER" >/dev/null 2>&1; then
|
|
echo "container not found: $CONTAINER" >&2
|
|
exit 1
|
|
fi
|
|
ARGS=(logs -f)
|
|
[ -n "$TAIL" ] && ARGS+=(--tail "$TAIL")
|
|
[ -n "$SINCE" ] && ARGS+=(--since "$SINCE")
|
|
docker "${ARGS[@]}" "$CONTAINER" 2>&1
|
|
fi | gawk '
|
|
function fmt_pct(v, got) {
|
|
return got ? sprintf("%5.1f%%", v) : " n/a"
|
|
}
|
|
|
|
function flush_pending() {
|
|
if (!have) return
|
|
printf "\033[37m[%s]\033[0m \033[36mPP\033[0m: \033[32m%6.1f t/s\033[0m | \033[31mTG\033[0m: \033[32m%6.1f t/s\033[0m | \033[33mK/V Cache\033[0m: \033[36m%6s\033[0m | \033[35mPfx Cache\033[0m: \033[36m%5.1f%%\033[0m | \033[34mDraft Acc\033[0m: \033[36m%6s\033[0m\n", \
|
|
pending_ts, pending_pp, pending_tg, fmt_pct(pending_kv, got_kv), pending_pfx, fmt_pct(pending_draft, got_draft)
|
|
fflush()
|
|
have = 0
|
|
}
|
|
|
|
/Avg prompt throughput:/ {
|
|
flush_pending()
|
|
if (match($0, /INFO ([0-9-]{5} [0-9:]{8})/, t)) pending_ts = t[1]
|
|
match($0, /Avg prompt throughput: ([0-9.]+)/, p); pending_pp = p[1] + 0
|
|
match($0, /Avg generation throughput: ([0-9.]+)/, g); pending_tg = g[1] + 0
|
|
if (match($0, /Prefix cache hit rate: ([0-9.]+)/, c)) pending_pfx = c[1] + 0
|
|
if (match($0, /KV cache usage: ([0-9.]+)/, k)) {
|
|
pending_kv = k[1] + 0
|
|
got_kv = 1
|
|
} else {
|
|
got_kv = 0
|
|
}
|
|
if (match($0, /Avg Draft acceptance rate: ([0-9.]+)/, d)) {
|
|
pending_draft = d[1] + 0
|
|
got_draft = 1
|
|
} else {
|
|
got_draft = 0
|
|
}
|
|
have = 1
|
|
next
|
|
}
|
|
|
|
/Avg Draft acceptance rate:/ {
|
|
if (!have) next
|
|
if (match($0, /INFO ([0-9-]{5} [0-9:]{8})/, t) && t[1] != pending_ts) next
|
|
match($0, /Avg Draft acceptance rate: ([0-9.]+)/, d)
|
|
pending_draft = d[1] + 0
|
|
got_draft = 1
|
|
flush_pending()
|
|
}
|
|
|
|
END { flush_pending() }
|
|
'
|