Reduce the image processing latency in VLM (#11541)

Co-authored-by: qiuxuan.lzw <qiuxuan.lzw@alibaba-inc.com>
2025-10-16 15:00:03 -07:00
parent b0d1d717e1
commit fd389df96e
3 changed files with 16 additions and 3 deletions
@@ -221,6 +221,10 @@ class Envs:
    SGLANG_TRITON_PREFILL_TRUNCATION_ALIGN_SIZE = EnvInt(4096)
    SGLANG_TRITON_DECODE_SPLIT_TILE_SIZE = EnvInt(256)

+    # VLM
+    SGLANG_IMAGE_MAX_PIXELS = EnvInt(16384 * 28 * 28)
+    SGLANG_RESIZE_RESAMPLE = EnvStr("")
+
    # fmt: on