- 专利标题: ELECTRO-PHOTONIC NETWORK FOR MACHINE LEARNING
-
申请号: US18590689申请日: 2024-02-28
-
公开(公告)号: US20240201436A1公开(公告)日: 2024-06-20
- 发明人: Philip Winterbottom , Martinus Bos
- 申请人: Celestial AI Inc.
- 申请人地址: US CA Santa Clara
- 专利权人: Celestial AI Inc.
- 当前专利权人: Celestial AI Inc.
- 当前专利权人地址: US CA Santa Clara
- 主分类号: G02B6/12
- IPC分类号: G02B6/12 ; G02B6/30 ; G02B6/43 ; G06F7/50 ; G06F7/53 ; G06F7/544 ; G06N3/04 ; G06N3/067
摘要:
Various embodiments provide for electro-photonic networks, including a plurality of processing elements connected by bidirectional photonic channels, suited for implementing neural-network models. Weights of the model may be preloaded into memory of the processing elements based on assignments of neural nodes to processing elements implementing them, and routers of the processing elements can be configured to stream activations between the processing elements based on a predetermined flow of activations in the model.
信息查询