Cerebras 与 AWS 合作,CS-3 将部署到 AWS 数据中心并通过 Bedrock 提供推理
Cerebras 宣布 AWS 将在其数据中心部署 Cerebras CS-3 系统,并通过 AWS Bedrock 提供推理服务,运行领先的开源 LLM 和 Amazon Nova 模型。双方还在合作一种解耦架构,把 AWS Trainium 与 Cerebras WSE 配对,在相同硬件占用下提供 5 倍的高速 token 容量。该架构中 Trainium 负责 prefill 并计算 KV cache,经 Amazon EFA 传给 WSE 专做 decode,Cerebras 称其可每秒生成数千个输出 token,而 GPU 为数百个。AWS 与 Cerebras 将同时支持聚合与解耦两种配置,预计多数客户会按负载在两者间路由。
lt is an urgent and growing need for more fast inference across the industry. Cerebras is the market leader in high-speed AI inference, powering models from OpenAI, Cognition, and Meta at up to 3,000 tokens per second. Today's announcement brings that speed to AWS's global customer base. And through our unique collaboration on disaggregated inference, we'll be able to provide 5x more capacity for high-speed inference. Every time you ask AI a question, two distinct kinds of computation happen: pr
匹配字段正文
