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 12, 08:30 UTC
[Doc]: fixing typos in various files. (#29717)
Signed-off-by: Didier Durand <durand.didier@gmail.com>
Didier Durand committed
Nov 29, 2025 at 10:15 UTC
04a797cd0e94567ee2b158a0246d6112a197fdc5
5 files changed
+5
-5
vllm/distributed/device_communicators/pynccl_allocator.py
+1
-1
@@ -157,7 +157,7 @@ class nccl_symm_mem_context:
157
if self.disabled:
158
return self
159
assert self.pynccl_comm is not None, (
160
- "Symmetric memory requires pynccl to be initalized"
160
+ "Symmetric memory requires pynccl to be initialized"
161
)
162
assert self.pynccl_comm.nccl_version >= 22703, (
163
"NCCL version 2.27.3 or higher is required for NCCL symmetric memory"
vllm/distributed/parallel_state.py
+1
-1
@@ -1583,7 +1583,7 @@ def destroy_distributed_environment():
1583
1584
1585
def cleanup_dist_env_and_memory(shutdown_ray: bool = False):
1586
- # Ensure all objects are not freezed before cleanup
1586
+ # Ensure all objects are not frozen before cleanup
1587
gc.unfreeze()
1588
1589
destroy_model_parallel()
vllm/entrypoints/openai/serving_models.py
+1
-1
@@ -150,7 +150,7 @@ class OpenAIServingModels:
150
lora_request.base_model_name = base_model_name
151
152
# Validate that the adapter can be loaded into the engine
153
- # This will also pre-load it for incoming requests
153
+ # This will also preload it for incoming requests
154
try:
155
await self.engine_client.add_lora(lora_request)
156
except Exception as e:
vllm/model_executor/layers/quantization/quark/schemes/quark_ocp_mx.py
+1
-1
@@ -37,7 +37,7 @@ logger = init_logger(__name__)
37
# use `rocm_aiter_ops.is_asm_fp4_gemm_dynamic_quant_enabled()`
38
# for envs checks which does not require @cache anymore.
39
# triton kernel is torch compile compatible.
40
-# does not require direct registeration.
40
+# does not require direct registration.
41
# use `rocm_aiter_ops.triton_fp4_gemm_dynamic_qaunt`.
42
@cache
43
def is_rocm_aiter_fp4_asm_gemm_enabled() -> bool:
vllm/transformers_utils/repo_utils.py
+1
-1
@@ -171,7 +171,7 @@ def file_or_path_exists(
171
repo_id=model, filename=config_name, revision=revision
172
)
173
if isinstance(cached_filepath, str):
174
- # The config file exists in cache- we can continue trying to load
174
+ # The config file exists in cache - we can continue trying to load
175
return True
176
177
# NB: file_exists will only check for the existence of the config file on