From f5bdf5643f62172416bda6002a6ba2c36f50b7bf Mon Sep 17 00:00:00 2001 From: wassname <1103714+wassname@users.noreply.github.com> Date: Fri, 18 Sep 2026 18:25:30 +0800 Subject: [PATCH] Exclude int8 from WVS routing Signed-off-by: PI[gpt-5.4] --- scripts/wvs_score_all_options_refresh.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/scripts/wvs_score_all_options_refresh.py b/scripts/wvs_score_all_options_refresh.py index d1f79e6..35788b8 100644 --- a/scripts/wvs_score_all_options_refresh.py +++ b/scripts/wvs_score_all_options_refresh.py @@ -26,7 +26,7 @@ DENSE_BASELINE_USD = Decimal("3.5908606723") OSS_PROVIDER = { "allow_fallbacks": True, "require_parameters": True, - "quantizations": ["fp8", "int8", "bf16", "fp16"], + "quantizations": ["fp8", "bf16", "fp16"], } LANES = ("openai", "google", "xai", "muse", "kimi", "glm", "deepseek", "qwen") SPECIALIZED = ("batch", "free", "-pro", "-fast", "vision", "-vl", "-5v", "-4.6v", "-4.5v", "-code", "-codex", "coder", "audio", "clip", "image", "guard", "safeguard", "multi-agent", "embedding", "rerank")