Qianfan-VL: A Milestone Achievement in Chinese Multimodal AI with Domestic Chips
•
9
AI & ML interests
None defined yet.
Recent Activity
View all activity
Papers
Less Noise, More Voice: Reinforcement Learning for Reasoning via Instruction Purification
LoongFlow: Directed Evolutionary Search via a Cognitive Plan-Execute-Summarize Paradigm
Articles
collection of ERNIE 4.5 models.
-
baidu/ERNIE-4.5-VL-28B-A3B-Thinking
Image-Text-to-Text • 30B • Updated • 1.3k • 518 -
baidu/ERNIE-4.5-21B-A3B-Thinking
Text Generation • 22B • Updated • 320 • • 772 -
baidu/ERNIE-4.5-VL-424B-A47B-Base-Paddle
Image-Text-to-Text • 424B • Updated • 19 • 61 -
baidu/ERNIE-4.5-VL-424B-A47B-Base-PT
Image-Text-to-Text • 424B • Updated • 97 • • 76
collection of ERNIE 4.5 models.
-
baidu/ERNIE-4.5-VL-28B-A3B-Thinking
Image-Text-to-Text • 30B • Updated • 1.3k • 518 -
baidu/ERNIE-4.5-21B-A3B-Thinking
Text Generation • 22B • Updated • 320 • • 772 -
baidu/ERNIE-4.5-VL-424B-A47B-Base-Paddle
Image-Text-to-Text • 424B • Updated • 19 • 61 -
baidu/ERNIE-4.5-VL-424B-A47B-Base-PT
Image-Text-to-Text • 424B • Updated • 97 • • 76
Qianfan-vl model series. The models are mainly domain enhanced vision language model, targeting enterprise level multi modal understanding scenarios.