Higher priority for user input of max_prefill_tokens & format (#540)

This commit is contained in:
Ying Sheng
2024-06-12 21:48:40 -07:00
committed by GitHub
parent 1374334d38
commit fb9296f0ed
50 changed files with 817 additions and 569 deletions

View File

@@ -24,8 +24,8 @@ from vllm.model_executor.layers.vocab_parallel_embedding import (
ParallelLMHead,
VocabParallelEmbedding,
)
from vllm.model_executor.utils import set_weight_attrs
from vllm.model_executor.model_loader.weight_utils import default_weight_loader
from vllm.model_executor.utils import set_weight_attrs
from vllm.transformers_utils.configs.dbrx import DbrxConfig
from sglang.srt.layers.logits_processor import LogitsProcessor