1 / 3
§ 002Description
INN inside™ 模型基础版，面向 BIE 系列智算体运行环境
§ 003Inquire
§ 004Specifications
下载敬请期待
Training Speed (3 nodes)2,160,000 tokens/s
Inference Speed (3 nodes)12,600,000 tokens/s
Inference Latency小于100ms
High Load Stability长时间复杂任务 < 70°C
Benchmark: Kaggle Diabetes准确率 89.7%
Benchmark: Kaggle Heart Disease准确率 98.6%
Benchmark: Kaggle MNIST准确率 98.2%
Benchmark: 双螺旋准确率 99.5%
§ 005Classification
Product form
software
Scenarios
ai trainingai inference
CPU type
INTEL-XEON