Show HN: Inferbench, collect/share datapoints on GPU's inference performance (www.inferbench.com)

InferBench has been launched as an open, community-driven platform for collecting and sharing benchmark results related to local large language model (LLM) inference performance on various GPUs. Users can submit their own benchmarking data, which is then validated by the community, fostering a collaborative environment aimed at enhancing the quality of data on hardware performance. This initiative empowers researchers and developers to better understand their hardware capabilities and optimize model inference, a critical aspect in the deployment of AI applications. The significance of InferBench lies in its potential to provide a comprehensive and trustworthy leaderboard for GPU inference performance, addressing a crucial need in the AI/ML community. As LLMs become increasingly prevalent, understanding the inference efficiency on different hardware configurations is essential for both benchmarking and performance tuning. By enabling users to flag suspicious entries, InferBench also prioritizes data integrity, fostering a more reliable resource for developers seeking to make informed hardware choices when running AI models.