None
DE
Baidu open-sources ERNIE 4.5 series models, including multimodal MoE architecture
['Technode Feed', '.Wp-Block-Co-Authors-Plus-Coauthors.Is-Layout-Flow', 'Class', 'Wp-Block-Co-Authors-Plus', 'Display Inline', '.Wp-Block-Co-Authors-Plus-Avatar', 'Where Img', 'Height Auto Max-Width', 'Vertical-Align Bottom .Wp-Block-Co-Authors-Plus-Coauthors.Is-Layout-Flow .Wp-Block-Co-Authors-Plus-Avatar', 'Vertical-Align Middle .Wp-Block-Co-Authors-Plus-Avatar Is .Alignleft .Alignright']
TechNode
Baidu has officially open-sourced its ERNIE 4.5 series, releasing 10 models on Hugging Face, GitHub and its own PaddlePaddle ecosystem.
The lineup includes large-scale MoE (mixture of experts) models with activated parameter sizes of 47 billion and 3 billion (total parameters up to 424 billion), plus smaller dense models at 0.3 billion parameters.
A key feature is a multimodal heterogeneous MoE design that shares parameters across modalities while reserving dedicated spaces, aimed at enhancing tasks like vision-language reasoning without sacrificing text performance.
The models were trained and optimized using PaddlePaddle, with Baidu reporting MFU (model FLOPs utilization) of up to 47%.
[Baidu, in Chinese]Related
['ernie'
'paddlepaddle'
'47'
'including'
'visionlanguage'
'utilization'
'multimodal'
'billion'
'opensources'
'series'
'parameters'
'architecture'
'technode'
'weights'
'moe'
'models'
'baidu']