百度发布了ERNIE-5.5-Flash-Base,一个拥有2800亿总参数和每token 140亿激活参数的大型开源混合专家(MoE)模型。该模型在33.5万亿token上进行了预训练,支持100万token的上下文窗口用于微调和推理。百度发布了原始的预训练检查点,没有进行进一步微调,允许开发者将其适配到各种应用中。该模型的架构强调效率,拥有专门的注意力机制和优化的MoE层,旨在降低延迟和内存使用,尤其是在现代GPU上。
AI
Baidu released ERNIE-5.5-Flash-Base, a raw pretraining checkpoint with 280B total and 14B active parameters, 128K training context and 1M-token support, for further training.
<p>Baidu has released ERNIE 5.5 Flash Base, a 280B sparse MoE base model with 14B active parameters, trained on 33.5T tokens and published before any mid- or post-training.</p> <p>The post <a href="https://www.marktechpost.com/2026/11/05/baidu-ernie-5-5-flash-base-280b-moe/">Baid…
<h2> TL;DR </h2> <p>Baidu published ERNIE-5.5-Flash-Base on Hugging Face on November 5: a 280B-parameter mixture-of-experts base model with 14B active parameters per token, trained on 33.5T tokens. The FP8 repository is 288.4 GB (measured from the file listing), so it fits on eig…
<p>Nothing here is sponsored — no affiliate links on this page. This breaking article is not the daily buyer-intent post. Facts below come from the sources listed at the end, checked 5 November 2026. We have not tested ERNIE 5.5 Flash Base hands-on, and nothing here is a hands-on…
dev.to — LLM tag
TIER_1English(EN)·Mariano Gobea Alcoba·
<h2> Architectural Analysis of ERNIE 5.5 Flash Base: Optimizing Inference Latency through Model Distillation and Kernel Fusing </h2> <p>The release of ERNIE 5.5 Flash Base marks a significant iteration in Baidu’s large language model (LLM) strategy, shifting the paradigm from pur…
Baidu has open-sourced ERNIE-5.5-Flash-Base, the raw pretraining checkpoint of ERNIE 5.5 Flash. The sparse mixture-of-experts model has 280 billion total parameters, activates 14 billion per token and was trained on 33.5 trillion tokens, with inference context of up to 1 million …