Fix: use cu129 nightly variant (has x86_64 wheels)

glm-ocr.py

@@ -11,7 +11,7 @@
 # ]
 #
 # [[tool.uv.index]]
-# url = "https://wheels.vllm.ai/nightly"
+# url = "https://wheels.vllm.ai/nightly/cu129"
 #
 # [tool.uv]
 # prerelease = "allow"
@@ -25,10 +25,9 @@ GLM-OCR is a compact 0.9B parameter OCR model achieving 94.62% on OmniDocBench V
 Uses CogViT visual encoder with GLM-0.5B language decoder and Multi-Token Prediction
 (MTP) loss for fast, accurate document parsing.
 
-NOTE: Requires vLLM nightly wheels (GLM-OCR added in v0.16.0, PR #33005) and
-transformers>=5.1.0 (GLM-OCR support landed in stable release).
-As of 2026-02-13, vLLM nightly only has ARM wheels — x86_64 builds broken.
-Check https://wheels.vllm.ai/nightly for updated builds before running.
+NOTE: Requires vLLM nightly wheels from cu129 variant (GLM-OCR added in v0.16.0,
+PR #33005) and transformers>=5.1.0 (GLM-OCR support landed in stable release).
+Uses https://wheels.vllm.ai/nightly/cu129 which has x86_64 wheels.
 First run may take a few minutes to download and install dependencies.
 
 Features: