SenseTime Launches 'Office Raccoon' Mini-Program

TapTechNews June 24th news, SenseTime Technology announced that the data analysis product Office Raccoon developed based on SenseTime's SenseNova every day large model capability was officially launched in the mini-program version today - Raccoon Smart Assistant.

The official said that in the mini-program, files directly inserted in WeChat chat (xls/xlsx/csv/txt/json, etc.) can be refined by the smart assistant and various kinds of charts can be generated. All results are output in a streaming fashion, what you see is what you get, and can be interrupted at any time.

Office Raccoon is a native data analysis product of large models based on SenseTime's large language model, which can automatically convert data into meaningful analysis and visualization results through the user's natural language input. According to TapTechNews' previous report, SenseTime released the SenseNova 5.0 large model in April, which adopted the MOE Hybrid Expert Architecture and significantly improved in knowledge, mathematics, reasoning, and code capabilities.

The model is trained based on more than 10 terabytes of tokens and has a 200,000 inference context window (equivalent to about 365,000 Chinese characters), and the inference context window reaches about 200,000, providing a variety of large models and capabilities such as natural language processing, image generation, automatic data annotation, and custom model training.