bug: 根据安装文档，无法拉取昆仑安装包内容

[Lokkisen]

Bug Description

根据官方安装文档，进行到昆仑相关软件包安装时，kunlun_ops无法拉取，提示404

Install kunlun_ops

uv pip install "https://baidu-kunlun-customer.su.bcebos.com/aiak/mimo/20260227/kunlun_ops-0.1.58+ee39020a-cp310-cp310-linux_x86_64.whl"

sen@sendeMacBook-Pro vllm-kunlun-whls % curl -v -fLO "https://baidu-kunlun-customer.su.bcebos.com/aiak/mimo/20260227/kunlun_ops-0.1.58+ee39020a-cp310-cp310-linux_x86_64.whl"

• The requested URL returned error: 404
• Closing connection
  curl: (22) The requested URL returned error: 404
  另外两个包可以正常拉取

Steps to Reproduce

直接官方命令拉取

Expected Behavior

No response

Additional Context

No response

[xyDong0223]

再试一下，可能是连接不稳定，我尝试后是可以的

[Lokkisen]

再试一下，可能是连接不稳定，我尝试后是可以的

您好，晚上尝试后能够正常拉取。但是启动模型提示报错，帮忙看看是什么问题？
OSError: /opt/vllm_kunlun/lib/python3.10/site-packages/kunlun_ops/libxops_blocks.so: undefined symbol: _ZN5baidu3xpu5xblas13moe_fc_fusionILm2E7float16fLm3ES3_fLm2ES3_fLm2ES3_fS3_fS3_ifS3_fEEiPNS0_3api7ContextERKNS1_13NdScaleTensorIXT_EKT0_KT1_EERKNS7_IXT2_EKT3_KT4_EERKNS7_IXT5_EKT6_KT7_EERKNS7_IXT8_ET9_KT10_EERKNS1_9MoeDescV4IT11_T12_T13_T14_EERKNS1_13MoeEpilogueV2IT15_T16_T17_EERKNS1_24MoeFusionMultiStreamDescE

安装步骤如下：
干净的wjie520/vllm_kunlun:uv_base
cd /workspace/vLLM-Kunlun-0.15.1-dev
uv pip uninstall vllm
uv pip uninstall vllm-kunlun
uv pip install vllm==0.15.1 --no-build-isolation --no-deps
uv pip install -r requirements.txt
python vllm_kunlun/patches/patch_torch251.py
python setup.py build
python setup.py install
替换eval_frame.py
安装PyTorch→ kunlun_ops → xspeedgate_ops → cocopod
执行 source setup_env.sh
拉起 python -m vllm.entrypoints.openai.api_server
--host 0.0.0.0
--port 8356
--model /workspace/models/Qwen3-30B-A3B-Thinking-2507
--gpu-memory-utilization 0.9
--trust-remote-code
--max-model-len 32768
--tensor-parallel-size 1
--dtype float16