search

dgyoo / pa3

Recent image representation as PA3 of the computer vision class.
7 stars 0 forks source link

question about vl_svmtrain #10

Closed 20155004 closed 8 years ago

20155004 commented 8 years ago

Same as vl_simplenn how to use it??

오류 발생: trainBinarySvm (line 36) [w, b] = vl_svmtrain(idx2desc,cid,lambda, [],epsilon,10/lambda, biasMultiplier, biasLearningRate,[], loss, solver);

there are so many options..

Jinwoo-Jeon commented 8 years ago

when using vl_svmtrain with learning parameters, you should use with 'parameter name' and parameter value like

[w, b] = vl_svmtrain(x, y, lambda, 'epsilon', epsilon, ... 'biasMultiplier', biasMultiplier, 'biasLearningRate', biasLearningRate, 'loss', loss, 'solver', solver);

wonyJ commented 8 years ago

(20153584) Also, second parameter 'y' is a vector of binary label(If each data has same class id with cid, then label is 1. Else, label is -1).

20155004 commented 8 years ago

So we need 67times svmtrain for each class. right?

Jinwoo-Jeon commented 8 years ago

Yes