diff --git a/configs/nvidia-master.yaml b/configs/nvidia-master.yaml index ecfd9bffc..db57797bd 100644 --- a/configs/nvidia-master.yaml +++ b/configs/nvidia-master.yaml @@ -1541,7 +1541,7 @@ dsr1-fp8-b200-trt-mtp: - { tp: 8, ep: 1, conc-start: 4, conc-end: 256, spec-decoding: mtp } dsr1-fp8-h200-sglang: - image: lmsysorg/sglang:v0.5.12-cu130 + image: lmsysorg/sglang:v0.5.19-cu130 model: deepseek-ai/DeepSeek-R1-0528 model-prefix: dsr1 runner: h200 diff --git a/perf-changelog.yaml b/perf-changelog.yaml index f69af977a..a48f48503 100644 --- a/perf-changelog.yaml +++ b/perf-changelog.yaml @@ -7286,3 +7286,9 @@ description: - "Use thinking-on golden synthetic AL 3.51 for five-token DSpark throughput; disable adaptive verification and retain real verification for evals" pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/2974 + +- config-keys: + - dsr1-fp8-h200-sglang + description: + - "Update the DeepSeek-R1-0528 FP8 H200 SGLang image from v0.5.12-cu130 to v0.5.19-cu130. Model, TP8 topology, 8k1k workload, concurrency grid, launch flags and evals are unchanged." + pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/3010