Two months have passed since the release of POINT 1.0.TencentThe launch of POINTS 1.5 was announced on December 14th.
1AI notes that POINTS 1.5 still follows the classic LLaVA architecture used in POINTS 1.0, consisting of a vision encoder, a projector, and a large language model.
According to the official introduction, this generation of POINTS model not only takes into account the efficiency-first idea insisted in POINTS 1.0, but also greatly enhances the performance of the model.
According to Tencent, POINTS1.5-7B tops the list of the world's top sub-10B open source models, surpassing industry-leading models such as Qwen2-VL, InternVL2 and MiniCPM-V-2.5.
In terms of real-world applications, POINTS1.5 performs well in several aspects such as OCR of complex scenes, reasoning ability, key information extraction, Latex formula extraction, math, image translation, object recognition, and so on.