公开(公告)号：US11769035B1

公开(公告)日：2023-09-26

申请号：US16219751

申请日：2018-12-13

Applicant: Amazon Technologies, Inc.

Inventor： Lai Wei ,  Hagay Lupesko ,  Anirudh Acharya ,  Ankit Khedia ,  Sandeep Krishnamurthy ,  Cheng-Che Lee ,  Kalyanee Shriram Chendke ,  Vandana Kannan ,  Roshani Nagmote

IPC: G06N3/04 ,  G06N3/08

CPC classification number: G06N3/04 ,  G06N3/08

Abstract: Techniques are described automatically determining runtime configurations used to execute recurrent neural networks (RNNs) for training or inference. One such configuration involves determining whether to execute an RNN in a looped, or “rolled,” execution pattern or in a non-looped, or “unrolled,” execution pattern. Execution of an RNN using a rolled execution pattern generally consumes less memory resources than execution using an unrolled execution pattern, whereas execution of an RNN using an unrolled execution pattern typically executes faster. The configuration choice thus involves a time-memory tradeoff that can significantly affect the performance of the RNN execution. This determination is made automatically by a machine learning (ML) runtime by analyzing various factors such as, for example, a type of RNN being executed, the network structure of the RNN, characteristics of the input data to the RNN, an amount of computing resources available, and so forth.

公开(公告)号：US20230368028A1

公开(公告)日：2023-11-16

申请号：US18217929

申请日：2023-07-03

Applicant: Amazon Technologies, Inc.

Inventor： Hagay Lupesko ,  Anirudh Acharya ,  Cheng-Che Lee ,  Lai Wei ,  Kalyanee Chendke ,  Ankit Khedia ,  Vandana Kannan ,  Sandeep Krishnamurthy ,  Roshani Nagmote

IPC: G06N3/082 ,  G06N20/00 ,  G06N20/20 ,  G06T7/10 ,  G06F16/55 ,  G06F18/214

CPC classification number: G06N3/082 ,  G06N20/00 ,  G06N20/20 ,  G06T7/10 ,  G06F16/55 ,  G06F18/214

Abstract: Features related to systems and methods for automated generation of a machine learning model based in part on a pretrained model are described. The pretrained model is used as a starting point to augment and retrain according to client specifications. The identification of an appropriate pretrained model is based on the client specifications such as model inputs, model outputs, and similarities between the data used to train the models.