PulseAugur
中
实时 19:28:14
English(EN) b12385: qwen2vl : apply FFN weight scales (#32241)

llama.cpp 修复 qwen2vl 模型输出错误

llama.cpp 项目发布了一个更新 (b12385),解决了 qwen2vl 模型的一个问题。此前,由于未应用 FFN 权重缩放,该模型产生了不正确的输出,导致在 CUDA 上困惑度很高,在 CPU 上出现 NaN。此修复确保了将缩放正确传递给 build_ffn,类似于 Qwen2 模型处理它们的方式。 AI

影响 确保 qwen2vl 在 llama.cpp 框架内的正确模型行为和性能。

排序理由 这是 llama.cpp 项目中特定模型的错误修复版本,并非新模型发布或重大研究。

在 llama.cpp — Releases 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →

llama.cpp 修复 qwen2vl 模型输出错误

本文如何被排名

Signal score
36 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
这是 llama.cpp 项目中特定模型的错误修复版本,并非新模型发布或重大研究。
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
model release, infra
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

报道来源 [1]

  1. llama.cpp — Releases TIER_1 English(EN) · am17an ·

    b12385: qwen2vl : 应用 FFN 权重缩放 (#32241)

    <p>The qwen2vl graph passed NULL for the ffn_up/gate/down scales, so the<br /> per-tensor NVFP4 scales loaded by the generic pass were never applied<br /> and the model produced garbage (PPL ~1e6 on CUDA, NaN on CPU).</p> <p>Pass the scales to build_ffn like qwen2 does.</p> <p>Fi…