PulseAugur
中
实时 19:55:12
English(EN) How to Run ERNIE 5.5 Flash Base Locally (8 H100s or 2 B300s, and It Won't Chat)

百度发布 ERNIE 5.5 Flash Base，需要大量硬件且使用受限

作者 PulseAugur 编辑部 · [1 个来源] ·

百度已在 Hugging Face 上发布了 ERNIE 5.5 Flash Base，这是一个拥有 2800 亿参数的开放权重模型。该模型是一个原始的预训练检查点，不包含开箱即用的聊天能力，需要进行微调才能用于对话。在本地运行该模型需要大量硬件，估计在 4 位量化下至少需要 144 GB 的 VRAM，而对于 1M 的上下文窗口则需要高达 244 GB。与据称在许多基准测试中优于 ERNIE 5.5 Flash Base 的 DeepSeek V4 Flash Base 的更宽松的 MIT 许可证不同，百度的许可证还限制其在欧盟、英国和韩国的使用。 AI

影响 本地部署需要大量硬件且有使用限制，可能影响采用。

排序理由 来自主要 AI 实验室（百度）的模型发布，包含技术细节和性能声明。[lever_c_demoted from frontier_release: ic=1 ai=1.0]

在 dev.to — LLM tag 阅读 →

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

百度发布 ERNIE 5.5 Flash Base，需要大量硬件且使用受限

本文如何被排名

Signal score
75 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Significant
来自主要 AI 实验室（百度）的模型发布，包含技术细节和性能声明。[lever_c_demoted from frontier_release: ic=1 ai=1.0]
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. dev.to — LLM tag TIER_1 English(EN) · Alan West ·

    How to Run ERNIE 5.5 Flash Base Locally (8 H100s or 2 B300s, and It Won't Chat)

    <p>ERNIE 5.5 Flash Base won't chat with you. Baidu's new 280B open-weight model on Hugging Face is the raw end-of-pre-training checkpoint, it's served through <code>/v1/completions</code>, the smaller of its two repos is a 288 GB download, and just holding the weights takes two 2…

相关实体

相关话题