｜开发者
首页
课程
论坛
资源
模型仓库
镜像
驱动和工具包
TX5
TX8
文档
登录/注册
最新发布
热门
精华
发帖
12-11
优化yolov5模型在芯片上的推理性能小tips
反量化最本质的原理是将定点结果乘以其对应的反量化系数即可得到浮点结果
152
评论
点赞
收藏
12-11
推理结果如何反量化为浮点
反量化最本质的原理是将定点结果乘以其对应的反量化系数即可得到浮点结果
60
评论
点赞
收藏
12-11
清微骑士工具链学习路径指南
在线课程
17
评论
点赞
收藏
邮箱反馈
邮箱地址：IService@Tsingmicro.com
留言反馈
'%3e%3cpath%20d='M17.5042%201.4143L16.6832%200.514297C16.5295%200.351083%2016.3439%200.22121%2016.1378%200.132741C15.9318%200.0442718%2015.7098%20-0.000905394%2015.4856%201.37483e-05C15.2993%203.35893e-05%2015.115%200.0389871%2014.9447%200.114376C14.7743%200.189766%2014.6215%200.299927%2014.4962%200.437797L5.75597%209.36192C5.74776%209.38231%205.74424%209.40428%205.74567%209.42621C5.74327%209.44348%205.74327%209.46101%205.74567%209.47829L4.82317%2013.0365C4.79584%2013.1395%204.79421%2013.2476%204.81842%2013.3514C4.84263%2013.4552%204.89195%2013.5515%204.96204%2013.6318C5.01216%2013.687%205.07315%2013.7312%205.14118%2013.7618C5.20921%2013.7923%205.28282%2013.8085%205.35739%2013.8092H5.50782L8.77419%2012.8385C8.80575%2012.8511%208.84097%2012.8511%208.87254%2012.8385L17.6032%203.92786C17.7504%203.7647%2017.8624%203.57294%2017.9322%203.36456C18.002%203.15618%2018.028%202.93564%2018.0088%202.71673C18.008%202.23424%2017.8282%201.76922%2017.5042%201.41173V1.4143ZM6.13204%2012.3268L6.60709%2010.4683L7.82465%2011.8254L6.13204%2012.3268ZM8.71439%2011.0841L7.3001%209.54129L13.5943%203.27858L15.0394%204.80278L8.71439%2011.0841ZM16.7501%202.88386L15.8803%203.82308L14.4853%202.30979L15.3364%201.38023C15.372%201.34244%2015.415%201.31233%2015.4626%201.29175C15.5103%201.27117%2015.5616%201.26056%2015.6135%201.26056C15.6654%201.26056%2015.7168%201.27117%2015.7644%201.29175C15.8121%201.31233%2015.855%201.34244%2015.8906%201.38023L16.7115%202.28023C16.7995%202.37353%2016.8491%202.49656%2016.8504%202.62479C16.8519%202.72144%2016.8165%202.81503%2016.7514%202.88645L16.7501%202.88386ZM16.7443%207.79079C16.6677%207.79232%2016.5922%207.80927%2016.5222%207.84064C16.4523%207.87201%2016.3894%207.91714%2016.3374%207.97336C16.2297%208.08907%2016.1713%208.24218%2016.1747%208.40021V16.0149C16.1805%2016.2149%2016.1073%2016.4092%2015.9711%2016.5559C15.8348%2016.7025%2015.6463%2016.7896%2015.4464%2016.7985H1.87247C1.67251%2016.7896%201.48404%2016.7025%201.34777%2016.5559C1.2115%2016.4093%201.13835%2016.2149%201.14411%2016.0149V2.61066C1.13835%202.41064%201.21152%202.21639%201.34781%202.06987C1.4841%201.92336%201.67256%201.83636%201.87247%201.82766H10.274C10.4212%201.8105%2010.5566%201.73867%2010.6534%201.6264C10.7502%201.51413%2010.8013%201.36961%2010.7966%201.22145C10.8026%201.07235%2010.7522%200.926464%2010.6552%200.81299C10.5583%200.699516%2010.4222%200.62682%2010.274%200.609447H1.87247C1.62358%200.602098%201.37592%200.647048%201.14548%200.741394C0.915047%200.835739%200.706968%200.977379%200.534692%201.15716C0.35849%201.34123%200.220861%201.55866%200.129893%201.79668C0.0389247%202.0347%20-0.00354637%202.2885%200.00497593%202.54316V16.0714C-0.0107201%2016.5694%200.168746%2017.0538%200.505108%2017.4214C0.668888%2017.6001%200.867356%2017.7436%201.08841%2017.8431C1.30946%2017.9425%201.54846%2017.9959%201.79082%2018H15.5422C15.7845%2017.9959%2016.0235%2017.9425%2016.2446%2017.8431C16.4656%2017.7436%2016.6641%2017.6001%2016.8279%2017.4214C17.1642%2017.0538%2017.3436%2016.5694%2017.328%2016.0714V8.38992C17.3306%208.23379%2017.2711%208.08302%2017.1626%207.97069C17.0542%207.85837%2016.9056%207.79367%2016.7494%207.79079H16.7443Z'%20fill='%23262727'/%3e%3c/g%3e%3cdefs%3e%3cclipPath%20id='clip0_3863_6406'%3e%3crect%20width='18'%20height='18'%20fill='white'/%3e%3c/clipPath%3e%3c/defs%3e%3c/svg%3e)
