Issues

• Add a Parallel Search MCP example for function calling

  alibaba/rtp-llm#1494 · Filter by author georgeatparallel opened on Oct 11, 2026

• RTP 受限解码（Tree Decode）的健壮性 bug：遇到 Trie 里不存在的 token 崩整个服务

  alibaba/rtp-llm#1450 · Filter by author LZYsmile opened on Sep 21, 2026
  1

• Add OrcaRouter as an optional AI provider

  alibaba/rtp-llm#1373 · Filter by author dulcestentaciones2920-debug opened on Sep 2, 2026

• Restore MLA smoke coverage for absorb_opt_len / decode_entrance / redundant_expert / hack_layer_num

  alibaba/rtp-llm#1342 · Filter by author netaddi opened on Aug 28, 2026

• [Docs] DingTalk and WeChat group QR codes in the README have expired

  alibaba/rtp-llm#1338 · Filter by author ZedingZhang opened on Aug 27, 2026

• Track production validation for topology KV candidate schedules

  alibaba/rtp-llm#1161 · Filter by author teerthsharma opened on Jul 7, 2026
  1

• Title: [Question] DeepSeek-V4 Support Plan

  alibaba/rtp-llm#1066 · Filter by author xulunfan opened on Jun 3, 2026
  1

• there are any performance test data for GLM 5.1?

  alibaba/rtp-llm#1064 · Filter by author dongyibo opened on Jun 2, 2026

• Consider using pre-built Flash Attention kernels via kernels

  alibaba/rtp-llm#1047 · Filter by author sayakpaul opened on May 28, 2026

• [Feature Request] Support for Hygon DCU Backend [功能申请] 海光 DCU (Hygon) 后端适配

  alibaba/rtp-llm#854 · Filter by author TracebaK opened on Apr 2, 2026
  1

• [BUG] pip install rtp.whl ERROR: is not a valid wheel filename.

  alibaba/rtp-llm#489 · Filter by author xueliangyang-oeuler opened on Dec 24, 2025
  3

• [CI/CD] Provide Dockerfile to build images from source

  alibaba/rtp-llm#487 · Filter by author xueliangyang-oeuler opened on Dec 24, 2025
  1

• [Feature] New Speculative Decoding Framework

  enhancement
  Task

  alibaba/rtp-llm#304 · Filter by author Vinkle-hzt opened on Oct 31, 2025
  

• Do you support Huawei Ascend NPU?

  alibaba/rtp-llm#281 · Filter by author baoojun opened on Oct 28, 2025
  2

• 注意力头维度为32会报错，有什么办法可以支持注意力维度为32的吗

  alibaba/rtp-llm#141 · Filter by author Abladol opened on Aug 8, 2025

• 支持在 H800上跑 int8 weight only 吗？

  alibaba/rtp-llm#140 · Filter by author hongqing1986 opened on Jul 14, 2025

• Why balance according to batch size in a technical report?

  alibaba/rtp-llm#139 · Filter by author nannaer opened on Jun 23, 2025

• 源码编译bazelisk build下载依赖失败，请问该如何操作

  alibaba/rtp-llm#138 · Filter by author liweiqing1997 opened on Jun 12, 2025

• git clone failed with error: git-lfs: command not found

  alibaba/rtp-llm#137 · Filter by author zhaohaidao opened on May 31, 2025

• socket.gaierror: [Errno -5] No address associated with hostname

  alibaba/rtp-llm#136 · Filter by author Greatpanc opened on Apr 15, 2025

• MODEL_TYPE 需要填写的是什么内容，参考文档中不是很清晰

  alibaba/rtp-llm#135 · Filter by author Greatpanc opened on Apr 15, 2025

• 最新代码源码编译报错

  alibaba/rtp-llm#132 · Filter by author leyiwang opened on Feb 24, 2025
  5

• what's the decode token speed of 7b qwen gptq int4 model on v100?

  alibaba/rtp-llm#131 · Filter by author FWXT opened on Feb 20, 2025

• Is there any tutorial document on how to use pd-disaggregation?

  alibaba/rtp-llm#130 · Filter by author Dimensionzw opened on Feb 7, 2025
  1

• support ebedding和reranker 部署时量化8

  alibaba/rtp-llm#124 · Filter by author invisifire opened on Nov 7, 2024
  3