Stack Exchange Network

Stack Exchange network consists of 175 Q&A communities including Stack Overflow, the largest, most trusted online community for developers to learn, share their knowledge, and build their careers.

Visit Stack Exchange

Questions tagged [transfer-learning]

The tag has no usage guidance.

0
votes
0answers
10 views

Domain Adaption with different tasks and domains

I know there exist plenty of deep learning algorithms for domain adaption (ADDA, DIRT-T, etc..), as long as the task keeps the same, e.g. I want to transfer knowledge from SVHN dataset to MNIST ...
0
votes
0answers
22 views

Can transfer learning be applied to predict sales

Let matrix A be a user item matrix.Upon performing UV decomposition , I get a user factor matrix and factor entity matrix. The company I am interning at doesn't keep track of the user factor matrix.I ...
2
votes
1answer
27 views

How to properly resize input images for transfer learning

I have to resize some images of different size to 224x224 before they can be passed as input for VGG19, and then apply ...
2
votes
1answer
116 views

Why is performance worse when my time-series data is not shuffled prior to a train/test split vs. when it is shuffled prior to the split?

We are running RandomForest model on a time-series data. The model is run in real time and is refit every time a new row is added. Since it is a timeseries data, we set shuffle to false while ...
1
vote
0answers
5 views

Why not mixing original training in favor of Pseudo-rehearsal?

In multiple task transfer training, I just learnt that Pseudo-rehearsal can be used to solve the problem of 'catastrophic forgetting' problem, i.e. NN forgets the original generalization while ...
1
vote
0answers
14 views

Supervisory information through side output in convolutional neural network

I am trying to implement this paper https://ieeexplore.ieee.org/document/7828014 Here they have mentioned text local (edge) and global regions as supervisory information. Side output is generated ...
1
vote
1answer
19 views

Transfer learning - small database

I am trying to use transfer learning in medical (ultrasound pictures). The problem is - I have very limited picture database = 400 (360+40). I am using resnet50 (I don't think this is important but ...
0
votes
0answers
18 views

Training InceptionV3/Resnet on custom data with CoreML/CreateML

I'm new to the CoreML/CreateML scene. I know that there is an image classifier built by Apple that can be easily trained by drag/dropping data, with pretty good accuracy. What I am wondering is, is ...
0
votes
1answer
17 views

Preprocessing for finetuned CNN model from pretrained models

Is it necessary to preprocess the images the same way as they were during the training of pretrained models in our finetuned model to use it for a different classification task ? Say, I have a ...
2
votes
0answers
35 views

How to help neuronal network with an other model

I am working on an image classification problem, the input data normally is images to classify, but I thought latitude and longitude would play something on these satellite images. I sorted by ...
0
votes
1answer
56 views

Why and how BERT can learn different attentions for each head?

https://towardsdatascience.com/deconstructing-bert-distilling-6-patterns-from-100-million-parameters-b49113672f77 I read the blog above. It visualizes that different color/head has different ...
1
vote
0answers
27 views

Pre training for multi label classification

I have to pre train a model in order to do transfer learning or fine tuning in multi label classification. I'm pretraining with cifar10 dataset and I wonder if I have to use for the pre training '...
0
votes
2answers
67 views

Does fine-tuning of transferred layers perform better than frozen transferred layers?

I recently learned concepts of transfer learning. Is it necessarily true that fine-tuning of transferred layers perform better than frozen transferred layer? why?
0
votes
1answer
39 views

Image features (produced by VGG19) do not properly train an ANN in Keras

I've used a VGG16 network to extract features from an image dataset, creating a features dataset using the following code: ...
2
votes
1answer
158 views

Which is the fastest image pretrained model?

I had been working with pre-trained models and was just curious to know the fastest forward propagating model of all the computer vision pre-trained models. I have been trying to achieve faster ...
1
vote
0answers
26 views

How to implement facial attendance system using less number of images of particulars

I have a project to implement facial Attendance where I have 5-6 images of particular and when individual comes, model should map the current image with person's earlier available images so if ...
1
vote
1answer
43 views

Resource and useful tips on Transfer Learning in NLP

I have a few label data for training and testing a DNN. Main purpose of my work is to train a model which can do a binary classification of text. And for this purpose, I have around 3000 label data ...
1
vote
1answer
19 views

Simple question about prediction classes of item in question vs not item in question

Let's say I wanted to use transfer learning to train a model to detect object A vs everything else. In this case, do I provide 2 types of input, images of object A and images of everything else, and ...
0
votes
1answer
174 views

Keras cNN Transfer Model: Reduce Final Model Size?

I'm working with multiple cNNs to be ran on mobile devices. If I create these cNNs from scratch (black n white, 256x256), I'm able to produce a binary classification model of about 10mb, which is a ...
1
vote
1answer
26 views

What tasks you train with one set of features and predict with another?

The most common scenario in supervised learning is to have data points with a set of features and train a model to make classification predictions afterward. Usually, for predictions to make sense ...
0
votes
1answer
41 views

correcting conditional and marginal distribution in transfer learning

I understand that in case of transfer learning, we can have the target and the source data having different domain distributions. In such cases, authors in many papers suggest bringing the marginal ...
0
votes
0answers
469 views

Intermediate layer output from pretrained TensorFlow model

I want to use the ResNet architecture to extract features from an image. That is, I plan to pass the image through ResNet, and take the feature maps from the final convolutional layer (the one just ...
3
votes
1answer
205 views

Transfer learning by concatenating the last classification layer

Before going into an obvious XY problem, I will explain you what I'm trying to do. I'm training a simple MobileNet pre-trained with Imagenet for multiclass classification. What I do is freeze all the ...
5
votes
2answers
121 views

Can neural networks be adapted without recreating them completely?

If I have, for example, a classification network which can tell if there is a dog or a cat in a picture, is it possible to adapt the network so it can also learn to detect a mouse? Without making a ...
2
votes
1answer
225 views

Fine-tuning NLP models

In computer vision, if we don't have a large training set, a common method is to start with a pre-trained model for some related task (e.g., ImageNet) and fine-tune that model to solve our problem. ...
3
votes
1answer
247 views

Keras bug NasNetlarge no top

I am trying to use NasNetlarge in Keras without the top but I cant get rid of the top: ...
1
vote
0answers
16 views

Domain adaption vs. heirarchical model - when to use which?

I know a little bit about domain adaption and also about random effects models, but I'm a little unsure if they're at all compatible. Domain adaptive models (based on ANNs) usually try to find some ...
1
vote
0answers
16 views

How to change the pooling to adaptpooling before FC layer in the Inception-ResnetV2 model in keras

Now I am using the keras model: Inception-ResnetV2 to do image classification using transfer learning. The main code about this model is as following: ...
2
votes
0answers
140 views

GTX 1080t ti rans out of memory

I have 60000 images divided into two classes. I have tried to build transfer learning with pretrained ResNet50 but my new GTX 1080 ti returns -1 after couples of epochs. My guess is that it runs out ...
3
votes
2answers
1k views

Is there any proven disadvantage of transfer learning for CNNs?

Suppose I know that I want to use a ResNet-101 architecture for my specific problem. There are ReseNet-101 models trained on ImageNet. Is there any disadvantage of using those pre-trained models and ...
2
votes
1answer
2k views

What are the consequences of not freezing layers in transfer learning?

I am trying to fine tune some code from a Kaggle kernel. The model uses pretrained VGG16 weights (via 'imagenet') for transfer learning. However, I notice there is no layer freezing of layers as is ...
1
vote
2answers
18 views

a simple way to test wether a tree-based classifier would transfer well to a target population?

I trained a tree-ensemble classifier (XGBOOST) on population A, validated it and I'm satisfied with its accuracy (AUC 0.78). Now I'm trying to transfer it to a slightly different population B, and ...
1
vote
0answers
563 views

Transfer learning: Poor performance with last layer replaced

I am using a transfer learning approach. For this I followed the tensorflow for poets tutorial. I use a pre-trained InceptionV3 architecture trained on the Imagenet dataset. The last layer and the ...
5
votes
2answers
2k views

Why is input preprocessing in VGG16 in Keras not 1/255.0

I am just trying to use pre-trained vgg16 to make prediction in Keras like this. ...
1
vote
0answers
65 views

Where to find pre-trained models for transfer learning [closed]

I am new to the machine learning field, but I wanted to try and implement a simple texture classification algorithm with Keras (to give a clear idea about what I mean as 'texture', [this][1] is the ...
2
votes
1answer
599 views

Transfer learning (on pre-trained inception net model) for multi label classification is giving similar probability for all labels

Number of labels: 1000, Dataset size: 200000 images Final probability for 1000 labels is in the range of 0.3 to 0.34. I was expecting large variation in probabilities. Can someone tell me what I am ...
5
votes
2answers
659 views

What is the purpose of untrainable weights in Keras

In this page it is mentioned that when trainable=false, the weight won't be updated and is used for optimization, too. But I still do not understand how it can be ...