Skip to content

GitLab

Unverified Commit 15310b51 authored by Michael Goin's avatar Michael Goin Committed by GitHub
Browse files

[Bugfix] Use LoadFormat values for `vllm serve --load-format` (#7784)

parent 57792ed4
Show whitespace changes
Inline Side-by-side
Showing with 3 additions and 6 deletions
vllm/engine/arg_utils.py
View file @ 15310b51
...@@ -9,8 +9,8 @@ import torch ...@@ -9,8 +9,8 @@ import torch
import vllm.envs as envs import vllm.envs as envs
from vllm.config import (CacheConfig, DecodingConfig, DeviceConfig, from vllm.config import (CacheConfig, DecodingConfig, DeviceConfig,
EngineConfig, LoadConfig, LoRAConfig, ModelConfig, EngineConfig, LoadConfig, LoadFormat, LoRAConfig,
ObservabilityConfig, ParallelConfig, ModelConfig, ObservabilityConfig, ParallelConfig,
PromptAdapterConfig, SchedulerConfig, PromptAdapterConfig, SchedulerConfig,
SpeculativeConfig, TokenizerPoolConfig) SpeculativeConfig, TokenizerPoolConfig)
from vllm.executor.executor_base import ExecutorBase from vllm.executor.executor_base import ExecutorBase
...@@ -214,10 +214,7 @@ class EngineArgs: ...@@ -214,10 +214,7 @@ class EngineArgs:
'--load-format', '--load-format',
type=str, type=str,
default=EngineArgs.load_format, default=EngineArgs.load_format,
choices=[ choices=[f.value for f in LoadFormat],
'auto', 'pt', 'safetensors', 'npcache', 'dummy', 'tensorizer',
'bitsandbytes'
],
help='The format of the model weights to load.\n\n' help='The format of the model weights to load.\n\n'
'* "auto" will try to load the weights in the safetensors format ' '* "auto" will try to load the weights in the safetensors format '
'and fall back to the pytorch bin format if safetensors format ' 'and fall back to the pytorch bin format if safetensors format '
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment