Pin OneComp serving plugin to immutable tag nex-serving-v1 (was mutable main); content identical, repro-stable

README.md

@@ -31,7 +31,7 @@ serves on a single workstation with **2×96 GB GPUs**.
 
 > ⚠️ **This checkpoint uses a custom mixed-bit MoE format and does NOT load with
 > stock vLLM / Transformers / llama.cpp.** Serving requires the mixed-bit MoE plugin
-> ([**mmzz164/OneCompression**](https://github.com/mmzz164/OneCompression/tree/main/vllm_plugins/gptq),
+> ([**mmzz164/OneCompression**](https://github.com/mmzz164/OneCompression/tree/nex-serving-v1/vllm_plugins/gptq),
 > MIT) — see *How to serve*. If you want a plug-and-play quant, use one of the standard
 > GGUF/AWQ builds instead.
 
@@ -110,10 +110,10 @@ This is a **custom mixed-bit MoE checkpoint**. Stock loaders cannot read it. To
    quantization/serving base.
 2. Add the **mixed-bit grouped fused-MoE serving plugin** (the part that actually loads
    and runs per-expert 2/3/4/8-bit MoE — not in upstream) from
-   **[mmzz164/OneCompression](https://github.com/mmzz164/OneCompression/tree/main/vllm_plugins/gptq)**:
+   **[mmzz164/OneCompression](https://github.com/mmzz164/OneCompression/tree/nex-serving-v1/vllm_plugins/gptq)**:
    `mixed_moe.py` (the `mixed_gptq` `FusedMoEMethodBase`), plus `grouped_moe.py` and
    `fused_dq_gemm.py` (the Triton kernels). See
-   [`MIXED_BIT_MOE_SERVING.md`](https://github.com/mmzz164/OneCompression/blob/main/vllm_plugins/gptq/MIXED_BIT_MOE_SERVING.md).
+   [`MIXED_BIT_MOE_SERVING.md`](https://github.com/mmzz164/OneCompression/blob/nex-serving-v1/vllm_plugins/gptq/MIXED_BIT_MOE_SERVING.md).
 3. Serve with vLLM using pipeline-parallel across your two GPUs and the `mixed_gptq`
    quantization, e.g. (illustrative):
    ```bash
@@ -160,7 +160,7 @@ model nex-agi/Nex-N2-Pro; only the weight precision was changed.
 - **Fujitsu Research** for [OneComp / OneCompression](https://github.com/FujitsuResearch/OneCompression)
   (MIT), including the AutoBit allocator and QEP.
 - Mixed-bit MoE serving kernels:
-  [mmzz164/OneCompression](https://github.com/mmzz164/OneCompression/tree/main/vllm_plugins/gptq) (MIT).
+  [mmzz164/OneCompression](https://github.com/mmzz164/OneCompression/tree/nex-serving-v1/vllm_plugins/gptq) (MIT).
 
 ## Citation