84e8326e61
ComfyUI serves image generation to Open WebUI on the Tart VM (ai.quineglobal.com). Models unload when idle on both sides so they share the 64 GB. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
41 lines
1.2 KiB
Python
41 lines
1.2 KiB
Python
#!/usr/bin/env python3
|
|
"""Unload ComfyUI's models once it has been idle for IDLE_MINUTES.
|
|
|
|
ComfyUI keeps models in memory until something else needs the room, and on
|
|
this 64 GB Mac that something is Ollama. Run every few minutes by
|
|
com.comfyui.idle-unload.plist; the counterpart of OLLAMA_KEEP_ALIVE=30m.
|
|
"""
|
|
import json
|
|
import time
|
|
import urllib.request
|
|
|
|
BASE = "http://127.0.0.1:8188"
|
|
IDLE_MINUTES = 15
|
|
|
|
|
|
def get(path):
|
|
with urllib.request.urlopen(BASE + path, timeout=10) as r:
|
|
return json.load(r)
|
|
|
|
|
|
try:
|
|
queue = get("/queue")
|
|
except OSError:
|
|
raise SystemExit(0) # ComfyUI not running
|
|
if queue["queue_running"] or queue["queue_pending"]:
|
|
raise SystemExit(0)
|
|
|
|
# Newest finish time (ms) of any job in the history.
|
|
last = 0
|
|
for job in get("/history?max_items=20").values():
|
|
for kind, data in job["status"].get("messages", []):
|
|
last = max(last, data.get("timestamp", 0))
|
|
if time.time() - last / 1000 < IDLE_MINUTES * 60:
|
|
raise SystemExit(0)
|
|
|
|
req = urllib.request.Request(
|
|
BASE + "/free",
|
|
data=json.dumps({"unload_models": True, "free_memory": True}).encode(),
|
|
headers={"Content-Type": "application/json"},
|
|
)
|
|
urllib.request.urlopen(req, timeout=10)
|