From 91688a234bf6fa6d4196494356b0fc0454517c99 Mon Sep 17 00:00:00 2001 From: Vuong Hoang Date: Thu, 18 Jun 2026 23:50:34 -0700 Subject: [PATCH] revert(litellm): remove mistral-medium-3.5 entry (GPU0 reverted to small-4 heretic) --- stacks/litellm/conf/config.yaml | 13 ------------- 1 file changed, 13 deletions(-) diff --git a/stacks/litellm/conf/config.yaml b/stacks/litellm/conf/config.yaml index ede137a..97d0e0d 100644 --- a/stacks/litellm/conf/config.yaml +++ b/stacks/litellm/conf/config.yaml @@ -95,19 +95,6 @@ model_list: model_info: mode: chat - # --- Mistral Medium 3.5 (RecViking NVFP4, vLLM nightly) — TEMPORARY speed-check - # tenant DISPLACING mistral-small-4 on GPU 0 (2026-06-19). Loads the HF-format - # NVFP4 directly on :8012; served-name mistral-medium-3.5. NOTE: while this is - # up, mistral-small-4 (above) points at a DOWN backend — revert is the - # mistral-small-4-heretic stack. --- - - model_name: mistral-medium-3.5 - litellm_params: - model: hosted_vllm/mistral-medium-3.5 - api_base: http://10.250.50.54:8012/v1 - api_key: os.environ/VLLM_API_KEY - model_info: - mode: chat - # --- Selene 1 Mini 8B (AtlaAI judge, FP8) — restored on GPU1 after the # llama-swap teardown (was the Q6_K GGUF in the swap zoo). vLLM dynamic fp8, # :8011. Explicit entry shadows the "*" wildcard (which used to reach it via