-
Notifications
You must be signed in to change notification settings - Fork 237
Pull requests: SemiAnalysisAI/InferenceX
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
CollectiveX: add NCCL EP as a fourth pluggable EP backend
#2331
opened Jul 25, 2026 by
Oseltamivir
Collaborator
Loading…
Proposal (Unofficial, not merge-ready): DeepSeek-V4-Flash 4xH200 SGLang MTP recipe / 提案（非官方，尚不可合并）：DeepSeek-V4-Flash 4×H200 SGLang MTP 配方
#2329
opened Jul 25, 2026 by
kfastino
Loading…
8 of 17 tasks
add b300 vllm agentX single-node minimaxm3 fp4 EAGLE-GQA MTP
full-sweep-fail-fast
#2328
opened Jul 24, 2026 by
hshrivastava-droid
Collaborator
Loading…
[AMD] dsv4-fp4-mi355x-atom: latest atom-dev image + opt recipe / DSv4-Pro FP4 MI355X ATOM：升级最新 atom-dev 镜像 + 优化配方
full-sweep-fail-fast
#2327
opened Jul 24, 2026 by
zufayu
Collaborator
Loading…
[AMD][AgentX] MINIMAX-M3 FP4 MI355X agentX vLLM disagg
#2326
opened Jul 24, 2026 by
charxwu
Collaborator
Loading…
feat: add Qwen3.5 FP8 GB200 Dynamo-SGLang MTP recipes / 新增 Qwen3.5 FP8 GB200 Dynamo-SGLang MTP 配方
full-sweep-enabled
#2324
opened Jul 24, 2026 by
RohitNagraj
Collaborator
Loading…
feat(power): validate single-node GPU energy metrics / 校验单节点 GPU 能耗指标
#2323
opened Jul 23, 2026 by
edwingao28
Collaborator
Loading…
[WIP] [NV] GB300 DSV4 AgentX Dynamo-SGlang: session-routing upgrade
full-sweep-enabled
NVIDIA
#2319
opened Jul 23, 2026 by
csahithi
Collaborator
Loading…
Add Qwen3.5 FP8 GB300 Dynamo-SGLang MTP recipes / 新增 Qwen3.5 FP8 GB300 Dynamo-SGLang MTP 配方
full-sweep-enabled
#2316
opened Jul 23, 2026 by
RohitNagraj
Collaborator
Loading…
Refresh MiniMax-M3 NVFP4 B300 8k1k disaggregated EAGLE Dynamo-vLLM / 刷新 MiniMax-M3 NVFP4 B300 8k1k 分离式 EAGLE Dynamo-vLLM 配置
non-canary-full-sweep-enabled
Run the full sweep without the canary gate (full search space, no trim)
#2314
opened Jul 23, 2026 by
jasonlizhengjian
Collaborator
Loading…
feat: add Qwen3.5 RTX 6000 Pro vLLM recipe / 新增 Qwen3.5 RTX 6000 Pro vLLM 基准配置
full-sweep-fail-fast
#2312
opened Jul 23, 2026 by
Oseltamivir
Collaborator
Loading…
[NV] Refresh MiniMax-M3 NVFP4 B300 8k1k disaggregated Dynamo-vLLM / 更新 MiniMax-M3 NVFP4 B300 8k1k 分离式 Dynamo-vLLM 配置
non-canary-full-sweep-enabled
Run the full sweep without the canary gate (full search space, no trim)
#2310
opened Jul 23, 2026 by
jasonlizhengjian
Collaborator
Loading…
[AMD][AgentX] DSv4 MI355X agentic disagg: add EAGLE/MTP speculative-decoding arm / DSv4 MI355X 智能体分离式推理：新增 EAGLE/MTP 投机解码分支
agentx
AgentX benchmarks, recipes, and infrastructure
AMD
full-sweep-enabled
#2309
opened Jul 23, 2026 by
ichbinblau
Collaborator
Loading…
8 of 10 tasks
[AMD][AgentX] DSv4 MI355X agentic disagg: add high-concurrency dots using DEP8
agentx
AgentX benchmarks, recipes, and infrastructure
AMD
#2308
opened Jul 23, 2026 by
ichbinblau
Collaborator
Loading…
3 of 4 tasks
feat: add MiniMax-M3 RTX 6000 Pro vLLM recipe / 新增 MiniMax-M3 RTX 6000 Pro vLLM 基准配置
full-sweep-fail-fast
#2306
opened Jul 23, 2026 by
Oseltamivir
Collaborator
Loading…
[AMD][dsv4] Add MTP variant of dsv4-fp4-mi355x-sglang-disagg
#2305
opened Jul 23, 2026 by
Oseltamivir
Collaborator
Loading…
[WIP] Test DeepSeek-V4 GB300 Dynamo AgentX recipes
full-sweep-enabled
#2302
opened Jul 22, 2026 by
ivanium
Collaborator
Loading…
[Klaud Cold] Add GLM-5.2 NVFP4 GB300 agentic disaggregated dynamo-sglang recipe / 新增 GLM-5.2 NVFP4 GB300 智能体分离式 dynamo-sglang 配置
full-sweep-fail-fast
#2299
opened Jul 22, 2026 by
Oseltamivir
Collaborator
Loading…
Restore DeepSeek-V4 tool calling for AgentX evals
agentx
AgentX benchmarks, recipes, and infrastructure
evals-only
Suppress throughput and run only eval jobs; combine with all-evals to expand selection
full-sweep-fail-fast
#2295
opened Jul 21, 2026 by
cquil11
Collaborator
Loading…
Fix GB300 DeepSeek-V4 AgentX worker metrics / 修复 GB300 DeepSeek-V4 AgentX worker 指标采集
full-sweep-enabled
#2291
opened Jul 20, 2026 by
cquil11
Collaborator
Loading…
Experiment with MiniMax-M2.7 EAGLE on B200 / 在 B200 上实验 MiniMax-M2.7 EAGLE
full-sweep-fail-fast
#2289
opened Jul 20, 2026 by
cquil11
Collaborator
Loading…
[AMD][AgentX] Add Kimi-K2.7 MI355X vLLM agentic MTP
agentx
AgentX benchmarks, recipes, and infrastructure
AMD
full-sweep-enabled
#2286
opened Jul 20, 2026 by
haic0
Collaborator
Loading…
[AMD] exclude mia1-p01-g09 from AMD multi-node Slurm jobs
#2283
opened Jul 19, 2026 by
seungrokj
Collaborator
Loading…
1 task
Previous Next
ProTip!
no:milestone will show everything without a milestone.