r：来自glmnet和插入符号的系数对于相同的lambda是不同的

I've read a few Q&As about this, but am still not sure I understand, why the coefficients from glmnet and caret models based on the same sample and the same hyper-parameters are slightly different. Would greatly appreciate an explanation!I've read a few Q&As about this, but am still n