5
Qwen
请问目前K100支持的最新的vllm镜像版本是多少
2025-08-15 15:18:16
163

使用vllm 0.8.5的镜像推理模型一直出现以下错误
/home/triton/lib/Dialect/TritonGPU/IR/LinearLayoutConversions.cpp:658: std::optionalmlir::triton::LinearLayout mlir::triton::gpu::AMDWmmaEncodingAttr::toLinearLayout(llvm::ArrayRef  <long int> ) const: Assertion  ver == 1 || ver == 2' failed.  

全部评论(2)
只看楼主
网
网友HhYALT
2楼
试试docker pull image.sourcefind.cn:5000/dcu/admin/base/custom:vllm0.8.4-ubuntu22.04-dtk25.04-rc7-das1.5-py3.10-20250429-dev-qwen3-only
2025/8/20
网
网友f2RTwE
楼主
1楼

python3: /home/triton/lib/Dialect/TritonGPU/IR/LinearLayoutConversions.cpp:658: std::optional<mlir::triton::LinearLayout> mlir::triton::gpu::AMDWmmaEncodingAttr::toLinearLayout(llvm::ArrayRef<long int>) const: Assertion `ver == 1 || ver == 2' failed.

python3: /home/triton/lib/Dialect/TritonGPU/IR/LinearLayoutConversions.cpp:658: std::optional<mlir::triton::LinearLayout> mlir::triton::gpu::AMDWmmaEncodingAttr::toLinearLayout(llvm::ArrayRef<long int>) const: Assertion `ver == 1 || ver == 2' failed.
2025/8/15
没有更多内容了
--
楼主
Level.1
0 帖子
楼主热帖
暂无数据