Recently, the international artificial intelligence summit Neurips was held in New Orleans, USA, in the neurips 2023 large language model efficiency challenge held at the same timePERCENT Technology won the NVIDIA A100 track championship by a significant margin of 8 points ahead of second place.
Neurips, together with ICML and ICLR, is known as the three top international conferences in the field of machine learning, representing the highest level of research in the field of artificial intelligence, and is a grand event that gathers top scholars, scientists and engineers from all over the world, and has been held for 37 sessions so far, which has played a vital role in promoting the development and application of artificial intelligence technology.
During the conference, Microsoft and Meta co-hosted the LLM Efficiency Challenge, aiming to explore low-cost and efficient LLM R&D and fine-tuning solutions, in order to promote the wider application of large models. The competition includes two tracks, NVIDIA A, and the competition system requires participating teams to use a designated large model within 24 hours, fine-tune on a single GPU, improve the accuracy of the model, and complete specific tasks such as multiple-choice Q&A scenarios and conversation chats in the logical reasoning type. The challenge attracted more than 100 top teams from industry and academia around the world. The 100 percent team won the championship this time, thanks to the company's years of technology accumulation in the fields of artificial intelligence and machine learning. In the field of large models, PERCENT Technology invested in R&D earlier, has the ability to self-develop tens of billions of large models, and focuses on the research and development of multi-language and multi-task unified large models and smart emergency applications in combination with the needs of vertical scenarios such as government affairs and international business. In the future, PERCENT Technology will continue to rely on its own industry knowledge to continuously explore the application of large models in vertical fields and empower customers with intelligent transformation.