# 1.5B Parameter Scale

The RWKV-7 model adopts a flash linear attention architecture and supports multilingual text generation tasks.

Qwen2.5-1.5B-Instruct is a 1.5 billion parameter Chinese instruction fine-tuned model based on the Qwen2.5 architecture, suitable for text generation and reasoning tasks.

A downstream task model based on Qwen2-1.5B, specializing in re-ranking tasks, excelling in Chinese medical Q&A and general text re-ranking tasks.

The largest open-source language model trained on Australian law, with 1.5 billion parameters, suitable for natural language processing tasks in the Australian legal domain.

DeBERTa V2 XXLarge is an improved BERT model based on disentangled attention and enhanced mask decoding, with 1.5 billion parameters, surpassing BERT and RoBERTa performance on multiple natural language understanding tasks