Issues

Issues

• [API-Inference] GLM-5.3-Flash returns HTTP 401 while Qwen3.8-Flash-Next works with the same token

  modelscope/modelscope#1850 · opened on Nov 11, 2026

• 创空间免费 16G CPU（platform/2v-cpu-16g-mem）重新部署后一直「部署中」，ossfs 挂载后应用进程从未启动

  modelscope/modelscope#1849 · opened on Nov 10, 2026 · In-Progress
  1

• 关于 model scope 的API 推理接口是否已支持了音频推理？

  modelscope/modelscope#1842 · opened on Nov 4, 2026
  3

• [API-Inference][modelscope.ai] Qwen3.8-Max Responses API returns HTTP 405 security block for inert code text

  modelscope/modelscope#1837 · opened on Oct 28, 2026

• GLM 5.3 Flash in API Inference continues to output "lock"

  modelscope/modelscope#1832 · opened on Oct 27, 2026

• Mirror of kernels-community/* stopped syncing, and has no kernel version branches

  modelscope/modelscope#1828 · opened on Oct 23, 2026
  3

• 创空间仓库不小心推了几个无用分支，发现无权限删除

  modelscope/modelscope#1827 · opened on Oct 22, 2026
  1

• 2560x1600分辨率电脑edge打开设计师(AIGC)模板样式个人主页，页面顶端抽动会让CPU/GPU占用率飙升

  modelscope/modelscope#1826 · opened on Oct 22, 2026
  3

• [Bug] torch_utils distributed helpers hard-code CUDA: init_dist defaults to nccl and broadcast/collectives build tensors with device='cuda' (breaks Ascend NPU / XPU / CPU-gloo)

  modelscope/modelscope#1824 · opened on Oct 18, 2026
  1

• MCP hosted deployment always fails with 'unexpected end of JSON input'

  modelscope/modelscope#1823 · opened on Oct 17, 2026
  1

• ModelScope绑定的阿里云PAI默认工作空间误删除问题

  modelscope/modelscope#1822 · opened on Oct 14, 2026
  2

• MCP 广场：外部 OAuth Streamable HTTP 服务配置后仍不可部署、未提供工具信息

  modelscope/modelscope#1809 · opened on Oct 5, 2026
  4

• [Feature Request] 首页搜索框尺寸过小、视觉层级不足，建议放大并置于首屏核心位置

  modelscope/modelscope#1803 · opened on Sep 22, 2026
  1

• 网站前端关于分支的管理

  modelscope/modelscope#1800 · opened on Sep 18, 2026
  2

• ModelScope于llama.cpp集成无法使用

  modelscope/modelscope#1658 · opened on Mar 27, 2026
  14