Muse-Glimmer-NVFP4

模型介绍

The vendor-recipe quantized checkpoint of **Muse Glimmer**, packed for serving with SGLang. This is the model vendor's own `fp_awq` mixed-precision recipe (`MIXED_PRECISION`): **NVFP4** (E2M1, group 16) on most projections, **MXFP8** (E4M3, group 32) on the higher-sensitivity set (`v_proj`, `down_proj`, `lm_head`), norms in bf16 — 18.3 GiB versus 52 GiB for the bf16 export. Grab it and serve it directly by repo id; no conversion step needed. - **Base model:** `meta-models/Muse-Glimmer-30B` (Muse Glimmer bf16 export) - **Format:** Safetensors, 5 shards, 18.34 GiB - **Recipe:** vendor `fp_awq` hand-off packed with the SGLang fork's `convert_fp_awq_to_hf.py` (default flags — the recipe as shipped); quantization map in `hf_quant_config.json` (`quant_algo: MIXED_PRECISION`) - **Language model:** 52 layers, hidden 6656, 32 Q / 2 KV heads, vocab 202,048 - **Text-only serving** is recommended (`--language-model-only`); the vision tower ships in the base export, not in this quantized checkpoint Notes: decode is memory-bound, so the speedup tracks the 52→18 GiB weight reduction. On GB10, quantized *prefill* is slower than bf16 at batch ≥ 4 (~650 vs ~1500 input tok/s) — decode-heavy workloads win, prefill-heavy workloads should measure. Correctness smoke-verified (greedy arithmetic and generation with correct stop tokens); full accuracy suite pending.

主要特点

特点摘要待依据来源补齐，不凭模型名称推断能力。

局限与注意事项

暂无足够来源支持具体缺点结论。请重点验证实际任务表现、时延、数据政策、许可与版本生命周期。

来源报价与费用分析

同一模型在不同渠道、地区和配置下可能采用不同价格。下表独立保留各渠道报价，不将渠道价格视为厂商官方统一价格。

公开评测与运行配置

暂无已采集的可核验公开评分。模型未上榜不代表能力较低。

常见问题

资料来源与可追溯记录

许可证：来源未披露，须进一步核验。模型介绍与第三方评测均以原来源的具体版本和方法为准。