摘要:
用Pytorch写一个神经网络的步骤: Prepare dataset Design model using Class (inherit from nn.Module) Construct loss and optimizer (using Pytorch API) Training cycle 阅读全文
摘要:
title: Graph-Based Social Relation Reasoning, 2020 task: we propose a simpler, faster, and more accurate method named graph relational reasoning netwo 阅读全文
摘要:
假设数据集是独立同分布的,可以将数据集划分为不同的比例:Train Set and Test Set. 同时在Train Set and Test Set上做精度测试,或者隔一段时间在Test Set上做测试,来判断训练模型是否发生过拟合,受否需要提前的终止,目的是选择最好的模型参数。(严格的说,其 阅读全文