From 69adaaf222ff72e824724a64ddbeffa5d5d4bfd1 Mon Sep 17 00:00:00 2001
From: Rick Staa <rick.staa@outlook.com>
Date: Wed, 29 Jul 2026 16:55:25 +0200
Subject: [PATCH] refactor(vllm): take the HF token as HF_TOKEN in .env

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---
 vllm/.env.example | 2 +-
 vllm/compose.yml  | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/vllm/.env.example b/vllm/.env.example
index 27a20ab..75aa41d 100644
--- a/vllm/.env.example
+++ b/vllm/.env.example
@@ -4,7 +4,7 @@
 # Model vLLM serves (must match the client's --model).
 VLLM_MODEL=Qwen/Qwen2.5-0.5B-Instruct
 # Only needed for gated HuggingFace models.
-HUGGING_FACE_HUB_TOKEN=
+HF_TOKEN=
 
 # --- On-chain (paid) only below; offchain ignores these. ---
 
diff --git a/vllm/compose.yml b/vllm/compose.yml
index c64c456..088a750 100644
--- a/vllm/compose.yml
+++ b/vllm/compose.yml
@@ -40,7 +40,7 @@ services:
       - "${VLLM_GPU_MEM_UTIL:-0.5}"
     environment:
       # Only needed for gated models; leave empty otherwise.
-      - HUGGING_FACE_HUB_TOKEN=${HUGGING_FACE_HUB_TOKEN:-}
+      - HUGGING_FACE_HUB_TOKEN=${HF_TOKEN:-}
     # vLLM needs an NVIDIA GPU.
     deploy:
       resources: