removed 5 sec sleep and warmup request on load

This commit is contained in:
Colter Downing
2025-10-28 15:28:30 -07:00
parent 9f5a432513
commit d190308329
+5 -5
View File
@@ -315,10 +315,10 @@ class Backend:
with open(BENCHMARK_INDICATOR_FILE, "r") as f: with open(BENCHMARK_INDICATOR_FILE, "r") as f:
log.debug("already ran benchmark") log.debug("already ran benchmark")
# trigger model load # trigger model load
payload = self.benchmark_handler.make_benchmark_payload() # payload = self.benchmark_handler.make_benchmark_payload()
_ = await self.__call_api( # _ = await self.__call_api(
handler=self.benchmark_handler, payload=payload # handler=self.benchmark_handler, payload=payload
) # )
return float(f.readline()) return float(f.readline())
except FileNotFoundError: except FileNotFoundError:
pass pass
@@ -393,7 +393,7 @@ class Backend:
) )
# some backends need a few seconds after logging successful startup before # some backends need a few seconds after logging successful startup before
# they can begin accepting requests # they can begin accepting requests
await sleep(5) # await sleep(5)
try: try:
max_throughput = await run_benchmark() max_throughput = await run_benchmark()
self.__start_healthcheck = True self.__start_healthcheck = True