From 2c1f34da5a078785a2541e3f7f9f822e84676b49 Mon Sep 17 00:00:00 2001 From: functionstackx <47992694+functionstackx@users.noreply.github.com> Date: Sun, 2 Aug 2026 23:04:58 -0400 Subject: [PATCH] feat(golden-al): add GLM-5.2 MTP golden AL distribution [skip-sweep] MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Golden acceptance-length curve for GLM-5.2 with MTP (B300, vLLM, SPEED-Bench coding category, thinking_on, temperature 1.0, top_p 0.95, output_len 4096), per num_speculative_tokens 1-8. Values are verbatim from the speedbench-al.yml artifact of Actions run 28058352479. Measured on the GLM-5.2 FP8 checkpoint; one curve is committed per model, and the NVFP4/MXFP4 checkpoints ship the same nextn head. 中文：新增 GLM-5.2 MTP 黄金接受长度(AL)分布（B300、vLLM、SPEED-Bench coding 类别、thinking_on、temperature 1.0、top_p 0.95、output_len 4096）， 覆盖 num_speculative_tokens 1-8。数值与 Actions run 28058352479 的 speedbench-al 产物逐字节一致。基于 GLM-5.2 FP8 模型测量；每个模型仅提交 一条曲线，NVFP4/MXFP4 量化使用同一 nextn 头。 Co-Authored-By: Claude Opus 5 (1M context) --- golden_al_distribution/README.md | 1 + golden_al_distribution/README_zh.md | 1 + golden_al_distribution/glm5.2_mtp.yaml | 19 +++++++++++++++++++ 3 files changed, 21 insertions(+) create mode 100644 golden_al_distribution/glm5.2_mtp.yaml diff --git a/golden_al_distribution/README.md b/golden_al_distribution/README.md index 93666c4fba..15dcb6d172 100644 --- a/golden_al_distribution/README.md +++ b/golden_al_distribution/README.md @@ -112,6 +112,7 @@ Before accepting an updated curve, reviewers should verify: | Qwen3.5 397B-A17B | MTP | [`qwen3.5_mtp.yaml`](qwen3.5_mtp.yaml) | [27317114007](https://github.com/SemiAnalysisAI/InferenceX/actions/runs/27317114007) | | Kimi K2.5 | EAGLE3 | [`kimik2.5_eagle3.yaml`](kimik2.5_eagle3.yaml) | [28122195822](https://github.com/SemiAnalysisAI/InferenceX/actions/runs/28122195822) | | MiniMax-M3 | EAGLE3 | [`minimaxm3_eagle3.yaml`](minimaxm3_eagle3.yaml) | [28061204145](https://github.com/SemiAnalysisAI/InferenceX/actions/runs/28061204145) | +| GLM-5.2 | MTP | [`glm5.2_mtp.yaml`](glm5.2_mtp.yaml) | [28058352479](https://github.com/SemiAnalysisAI/InferenceX/actions/runs/28058352479) | ## Primary references diff --git a/golden_al_distribution/README_zh.md b/golden_al_distribution/README_zh.md index e4231027e5..1285b613bd 100644 --- a/golden_al_distribution/README_zh.md +++ b/golden_al_distribution/README_zh.md @@ -112,6 +112,7 @@ gh workflow run speedbench-al.yml \ | Qwen3.5 397B-A17B | MTP | [`qwen3.5_mtp.yaml`](qwen3.5_mtp.yaml) | [27317114007](https://github.com/SemiAnalysisAI/InferenceX/actions/runs/27317114007) | | Kimi K2.5 | EAGLE3 | [`kimik2.5_eagle3.yaml`](kimik2.5_eagle3.yaml) | [28122195822](https://github.com/SemiAnalysisAI/InferenceX/actions/runs/28122195822) | | MiniMax-M3 | EAGLE3 | [`minimaxm3_eagle3.yaml`](minimaxm3_eagle3.yaml) | [28061204145](https://github.com/SemiAnalysisAI/InferenceX/actions/runs/28061204145) | +| GLM-5.2 | MTP | [`glm5.2_mtp.yaml`](glm5.2_mtp.yaml) | [28058352479](https://github.com/SemiAnalysisAI/InferenceX/actions/runs/28058352479) | ## 主要参考资料 diff --git a/golden_al_distribution/glm5.2_mtp.yaml b/golden_al_distribution/glm5.2_mtp.yaml new file mode 100644 index 0000000000..afebacac98 --- /dev/null +++ b/golden_al_distribution/glm5.2_mtp.yaml @@ -0,0 +1,19 @@ +# Source GitHub Actions run: https://github.com/SemiAnalysisAI/InferenceX/actions/runs/28058352479 +# Acceptance Length (AL) reference values measured with SPEED-Bench. +# dataset: coding | temperature: 1.0 | top_p: 0.95 | output_len: 4096 +# thinking_on chat_template_kwargs: {"enable_thinking": true} +# thinking_off chat_template_kwargs: {"enable_thinking": false} +# Measured on glm-5.2-fp8 (B300, vLLM MTP), per num_speculative_tokens. +# Auto-generated by benchmarks/single_node/speedbench/glm52_fp4_b300_vllm.sh (speedbench-al.yml). +# +# key = num_speculative_tokens (MTP level); value = golden AL +glm-5.2-fp8: + thinking_on: + 1: 1.84 + 2: 2.50 + 3: 2.99 + 4: 3.33 + 5: 3.61 + 6: 3.78 + 7: 3.91 + 8: 4.06