refactor

2025-06-12 09:41:12 +08:00
parent fee58d44e4
commit f16adb729e
4 changed files with 14 additions and 22 deletions
--- a/nanovllm/config.py
+++ b/nanovllm/config.py
@@ -8,7 +8,7 @@ class Config:
    max_num_batched_tokens: int = 32768
    max_num_seqs: int = 512
    max_model_len: int = 4096
-    gpu_memory_utilization: float = 0.95
+    gpu_memory_utilization: float = 0.9
    enforce_eager: bool = False
    hf_config: AutoConfig | None = None
    eos: int = -1