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
[CI] Fix Pre-commit Mypy Error (#26181)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
Wentao Ye committed
Oct 3, 2025 at 19:08 UTC
767cbb011dfa6e7685739dd29b68cde0727e459a
1 file changed
+3
-3
vllm/transformers_utils/config.py
+3
-3
@@ -515,17 +515,17 @@ def maybe_override_with_speculators(
515
from vllm.transformers_utils.configs.speculators.base import (
516
SpeculatorsConfig)
517
518
- vllm_speculative_config = SpeculatorsConfig.extract_vllm_speculative_config(
518
+ speculative_config = SpeculatorsConfig.extract_vllm_speculative_config(
519
config_dict=config_dict)
520
521
# Set the draft model to the speculators model
522
- vllm_speculative_config["model"] = model
522
+ speculative_config["model"] = model
523
524
# Override model and tokenizer with the verifier model from config
525
verifier_model = speculators_config["verifier"]["name_or_path"]
526
model = tokenizer = verifier_model
527
528
- return model, tokenizer, vllm_speculative_config
528
+ return model, tokenizer, speculative_config
529
530
531
def get_config(