combining two neural networks (net1 is trained & net2 is untrained) in one bigger network

Question

Abdelwahab Afifi 2020-1-10

1
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/499688-combining-two-neural-networks-net1-is-trained-net2-is-untrained-in-one-bigger-network

评论： Russdsp 2021-6-11

I have two shallow neural network models net1, net2.

I have a trained net1. Then I need to connect a net2 to a net1. Hence I treat the combined (net2+net1) as one large net and then train the big net . However, during the second training phase, we only need to update the weights corresponding to net2 (and keep net1 weights fixed). Is it possible in MATLAB? any suggested links or articles to perform this part.

2 个评论
显示无隐藏无

Xin Yee Tai 2020-7-14

Could you please share the code to build the neural network on the last photo? Thank you.

Russdsp 2021-6-11

Yes please share code. the green box in the middle doesnt seem like a layer- how is that there?

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Hiro Yoshino 2020-1-10

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/499688-combining-two-neural-networks-net1-is-trained-net2-is-untrained-in-one-bigger-network#answer_409497

How about this?:

https://www.mathworks.com/matlabcentral/answers/497749-avoid-training-certain-neurons#answer_407422?s_tid=prof_contriblnk

2 个评论
显示无隐藏无

Abdelwahab Afifi 2020-1-10

what about connecting two neural networks with each other?

Abdelwahab Afifi 2020-6-22

After training, how can I extract net2 from the big net ?

请先登录，再进行评论。

Answer 2

Greg Heath 2020-1-12

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/499688-combining-two-neural-networks-net1-is-trained-net2-is-untrained-in-one-bigger-network#answer_409750

Just

1. Save the outputs of net1 in a file

2. Use the file to train net2

Greg

2 个评论
显示无隐藏无

Beatriz Martins 2020-1-12

编辑：Beatriz Martins 2020-1-12

Greg Heath,

I am trying to save the output of a net1 and use it to train net2. However, during training of net1 the trainFasterRCNNObjectDetector function added new layers (initially I have 25 layers and the output of net1 has 35) and I get an invalid network. Do you know how can I solve this problem?

Abdelwahab Afifi 2020-6-20