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 14, 05:30 UTC
fix some typos (#24167)
Signed-off-by: co63oc <co63oc@users.noreply.github.com> Co-authored-by: Russell Bryant <rbryant@redhat.com>
co63oc committed
Sep 10, 2025 at 21:21 UTC
3144d90217081aa7b4d3367f0614eb13c35ad4ae
3 files changed
+3
-3
tests/v1/e2e/test_spec_decode.py
+1
-1
@@ -162,7 +162,7 @@ def test_eagle_correctness(
162
# TODO: Fix this flaky test
163
pytest.skip(
164
"TREE_ATTN is flaky in the test disable for now until it can be "
165
- "reolved (see https://github.com/vllm-project/vllm/issues/22922)")
165
+ "resolved (see https://github.com/vllm-project/vllm/issues/22922)")
166
167
# Generate test prompts inside the function instead of using fixture
168
test_prompts = get_test_prompts(mm_enabled)
vllm/entrypoints/openai/tool_parsers/internlm2_tool_parser.py
+1
-1
@@ -89,7 +89,7 @@ class Internlm2ToolParser(ToolParser):
89
try:
90
parsable_arr = action
91
92
- # tool calls are generated in an object in inernlm2
92
+ # tool calls are generated in an object in internlm2
93
# it's not support parallel tool calls
94
try:
95
tool_call_arr: dict = partial_json_parser.loads(
vllm/v1/sample/logits_processor/__init__.py
+1
-1
@@ -195,7 +195,7 @@ class AdapterLogitsProcessor(LogitsProcessor):
195
overridden in general. However, to implement custom constructor behavior -
196
especially any logic which operates on or stores `vllm_config`, `device`,
197
or `is_pin_memory` - `self.__init__(vllm_config, device, is_pin_memory)`
198
- must be overriden and the override must call
198
+ must be overridden and the override must call
199
`super().__init__(vllm_config, device, is_pin_memory)`
200
"""
201