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 8, 22:45 UTC
[compile] Remove runner type from ignored caching factor list. (#33712)
Signed-off-by: zhxchen17 <zhxchen17@fb.com>
Zhengxu Chen committed
Feb 4, 2026 at 05:56 UTC
a208439537a9071668b99dc8089db0dd8995034a
1 file changed
+2
-3
vllm/config/model.py
+2
-3
@@ -105,8 +105,8 @@ class ModelConfig:
105
content for `model_name` tag in metrics output when `served_model_name` is
106
not specified."""
107
model_weights: str = ""
108
- """Original model weights path. Used when the model is pulled from object
109
- storage (e.g., RunAI) to preserve the original URI while `model` points to
108
+ """Original model weights path. Used when the model is pulled from object
109
+ storage (e.g., RunAI) to preserve the original URI while `model` points to
110
the local directory."""
111
runner: RunnerOption = "auto"
112
"""The type of model runner to use. Each vLLM instance only supports one
@@ -324,7 +324,6 @@ class ModelConfig:
324
the final hidden states.
325
"""
326
ignored_factors = {
327
- "runner",
327
"convert",
328
"tokenizer",
329
"tokenizer_mode",