快讯
d1-omni-600M借Core ML在Apple silicon上运行,M5 Pro上每秒审核224条评论
Alex表示,Liquid AI开源的d1-omni-600M模型已可在Apple silicon上以Core ML运行:用GPU+神经引擎处理Civil Comments的5000条真实评论,在M5 Pro上达到每秒224条的审核速度,与人工评分者的一致率达95.4%,与PyTorch基准版持平。Liquid AI当日发布Open d1系列两个开源多模态模型:d1-3B(文本+视觉)和d1-omni-600M(文本+图像或文本+音频)。
事件来源
查看原文
RT Alex d1-omni-600M, now running on Apple silicon with Core ML ⚡ We pointed it at a firehose of 5,000 real comments from Civil Comments with GPU + ANE. moderating 224 comments/s on an M5 Pro. 95.4% agreement with human raters, same as the base PyTorch too. 🤗 Model: http://huggingface.co/FluidInference/d1-omni-600m-coreml 🛠️ SDK: http://github.com/FluidInference/FluidAudio Liquid AI: Today we release Open d1: two open-weight multimodal models in our d1 decision model family. > d1-3B: text + vision > d1-omni-600M: text + image or text + audio > Real-time decision making anywhere, from data centers such as @nvidia DGX to RTX workstations to Jetson at the edge.