4

hi im super new to this field(deep learning, computer vision) so this question may sound dumb.

In this link (https://github.com/GeorgeSeif/Semantic-Segmentation-Suite), there are pre-trained models (eg, ResNet101) called front end models. And they are used for Feature Extractor. I found these models are called backbone models/architectures generally. And the link says some of main models(eg. DeepLabV3, PSPNet) rely on pre-trained ResNet.
Also, Transfer Learning is to take a model trained on a large dataset and transfer its knowledge to a smaller dataset, right ?

Then my question is ,
1.Do the models that rely on pre-trained ResNet do transfer learning basically ?
2.if i use pretrained network like ResNet101 as backbone architecture of main model(like U-Net,SegNet) for image segmentation, is it considered as transfer learning ?

Sorry for my bad english,and i would highly appreiate if you answer this questin. Thank you.

Jon.O
  • 41
  • 2
  • 1
    1. unless its only used for resnet, then yeah. Trying to use another dataset/task taking advantage of already trained network in a similar task is transfer learning. 2. Yes, unlike the other cases where you probably only change last layer or two, here you add more but its same concept – juvian Jan 29 '20 at 05:06
  • 1
    This question is not about programming. Please ask at https://ai.stackexchange.com/. Thanks. – Mathias Müller Jan 29 '20 at 11:05
  • @juvian thanks for the reply! can I ask a question about the question 1 ? i read a paper PSPNet and there is no word " transfer learning" . so are these models that rely on pre-trained models considered as transfer learning in a 'broad sense' ? I mean i feel like the word 'transfer learning' is not defined strictly. Thank you though! – Jon.O Jan 29 '20 at 22:45
  • 2
    They mention they use a pretrained ResNet so they are applying transfer learning. Its just not mentioned as its implied by the pretrained part. I recommend reading [standford summary](http://cs231n.github.io/transfer-learning/) – juvian Jan 30 '20 at 02:27
  • 1
    @juvian thank you so much !!! i apreciate you :) – Jon.O Jan 30 '20 at 16:58

0 Answers0