A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 16, 20:45 UTC
[Misc] Remove abused noqa (#6619)
Woosuk Kwon committed
Jul 21, 2024 at 08:47 UTC
b6df37f9431b44e3e8802448c95cbc6989c3716e
1 file changed
+6
-5
vllm/distributed/device_communicators/shm_broadcast.py
+6
-5
@@ -110,7 +110,8 @@ class ShmRingBuffer:
110
lambda *args, **kwargs: None):
111
try:
112
self.shared_memory = shared_memory.SharedMemory(name=name)
113
- assert self.shared_memory.size == self.total_bytes_of_buffer # noqa
113
+ assert (
114
+ self.shared_memory.size == self.total_bytes_of_buffer)
115
except FileNotFoundError:
116
# we might deserialize the object in a different node
117
# in this case, this object is not used,
@@ -344,8 +345,8 @@ class MessageQueue:
345
time.sleep(RINGBUFFER_SLEEP_INTERVAL)
346
347
# if we wait for a long time, we should warn the user
347
- if time.monotonic(
348
- ) - start_time > VLLM_RINGBUFFER_WARNING_INTERVAL * n_warning: # noqa
348
+ if (time.monotonic() - start_time >
349
+ VLLM_RINGBUFFER_WARNING_INTERVAL * n_warning):
350
logger.warning(
351
"No available block found in %s second. ",
352
VLLM_RINGBUFFER_WARNING_INTERVAL)
@@ -398,8 +399,8 @@ class MessageQueue:
399
time.sleep(RINGBUFFER_SLEEP_INTERVAL)
400
401
# if we wait for a long time, we should warn the user
401
- if time.monotonic(
402
- ) - start_time > VLLM_RINGBUFFER_WARNING_INTERVAL * n_warning: # noqa
402
+ if (time.monotonic() - start_time >
403
+ VLLM_RINGBUFFER_WARNING_INTERVAL * n_warning):
404
logger.warning(
405
"No available block found in %s second. ",
406
VLLM_RINGBUFFER_WARNING_INTERVAL)