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
[BUG] Allow runai_streamer_sharded in config check (#26958)
Signed-off-by: ahao-anyscale <ahao@anyscale.com>
ahao-anyscale committed
Oct 15, 2025 at 20:05 UTC
582f2c6be7459a781d844e94f7511c198d756f2a
1 file changed
+5
-2
vllm/config/vllm.py
+5
-2
@@ -740,10 +740,13 @@ class VllmConfig:
740
"Overriding `load_format` to 'runai_streamer'"
741
)
742
self.load_config.load_format = "runai_streamer"
743
- elif self.load_config.load_format != "runai_streamer":
743
+ elif self.load_config.load_format not in (
744
+ "runai_streamer",
745
+ "runai_streamer_sharded",
746
+ ):
747
raise ValueError(
748
f"To load a model from S3, 'load_format' "
746
- f"must be 'runai_streamer', "
749
+ f"must be 'runai_streamer' or 'runai_streamer_sharded', "
750
f"but got '{self.load_config.load_format}'. "
751
f"Model: {self.model_config.model}"
752
)