Stack Overflow
Stack Overflow

Questions tagged [overfitting-underfitting]

143 questions
0
votes
1 answer

How should the weights for a WeightedRandomSampler be created for object detection when the labels are logits?

I've been trying to classify the VOC2007 dataset using CNN. The dataset has an imbalance, with multiple images having 'person' class and all the other classes have very few images and my model has been suffering from overfitting. In order to over…
Priya Ravichander
  • 83
  • 3
0
votes
1 answer

Catboost overfits training data but test performance increases

I'm training catboost on a dataset made of 41k observations and ~60 features. The dataset is a longitudinal series (9 years) that is spatially distributed. At the moment I'm just using random resampling of data, ignoring spatial and temporal…
Fran_civ
  • 31
  • 5
0
votes
1 answer

Is my model overfitting or underfitting ? .How to identify it from this loss curve?

I have created a model based on the mobileNet architecture to classify images. I have used 209222 images for training and 40000 images for validation. After I trained the model the loss curve looks like below. Is this model overfitting or…
Ranish
  • 3
  • 2
0
votes
1 answer

how to prevent overfitting/underfitting while using EfficientNet

i'm new in python so for my college final project, i use EfficientNet for the ripeness palm oil into 4 classification. i've total 5852 train data (1463 each) , and 132 test data (33 each). i used code from the internet. but the result is far from…
Muhammad Rifqi
  • 43
  • 8
0
votes
1 answer

Training loss is decreaseing but validation loss is cosntant. how to avoid overfitting

I want to retrain google's mediapipe hand landmark for more key-points detection but the model is only available in tflite format that can not be retrain. I created a model identical to mediapipe hand model, and I trained it with my custom data but…
Swapnil Gautam
  • 21
  • 3
0
votes
0 answers

Should i increase or decrease the loss when model start overfitting

I am doing image segmentation with a learning rate of 1e-4 and after 80,90 epochs I felt my model starts overfitting. This answer at StackOverflow suggests that small learning rate leads to overfitting and large learning rate act as a regularizer.…
Talha Anwar
  • 2,699
  • 4
  • 23
  • 62
0
votes
1 answer

Why randomforest max depth parameter's validation score not shrink, when overfitting occur

I made randomforest model, and visualized result. #training code from sklearn.datasets import load_digits from sklearn.ensemble import RandomForestClassifier from sklearn.model_selection import GridSearchCV digits = load_digits() forest_param =…
Luckydipper
  • 15
  • 3
0
votes
1 answer

Both validation loss and accuracy are increasing using a pre-trained VGG-16

So, I'm doing a 4 label x-ray images classification on around 12600 images: Class1:4000 Class2:3616 Class3:1345 Class4:4000 I'm using VGG-16 architecture pertained on the imageNet dataset with cross-entrpy and SGD and a batch size of 32 and a…
ahmed ben
  • 145
  • 1
  • 11
0
votes
2 answers

My validation loss is lower than my training loss, should I get rid of regularization?

I've heard a lot of people talk about some of the causes but they never really answer if it should be fixed or not. I checked my dataset for leaks and I took 20% for my validation set at random from a TFRecords dataset. I'm starting to suspect that…
PHO BOSS
  • 21
  • 1
  • 6
0
votes
3 answers

CNN + LSTM model for images performs poorly on validation data set

My training and loss curves look like below and yes, similar graphs have received comments like "Classic overfitting" and I get it. My model looks like below, input_shape_0 = keras.Input(shape=(3,100, 100, 1), name="img3") model =…
Vasanth Nag K V
  • 4,860
  • 5
  • 24
  • 48
0
votes
1 answer

Data Leakage during categorical variable handling?

I am fairly new to machine learning. I came across the concept of Data Leakage. The article says that always split the data before performing preprocessing steps. My question is, do steps such as discretization, grouping categories to a single…
Joe
  • 91
  • 6
0
votes
1 answer

Validation accuracy is getting lower than the training accuracy

This model is still training. And the Validation accuracy is getting lower than the training. This shows overfitting ? How I can overcome this ? I have used the MobileNet model. Can I do it by reducing the learning rate? Epoch 10/50 6539/6539…
user10960174
0
votes
1 answer

RNN/GRU Increasing validation loss but decreasing mean absolute error

I am new to deep learning and I try to implement an RNN (with 2 GRU layers). At first, the network seems to do it's job quite fine. However, I am currently trying to understand the loss and accuracy curve. I attached the pictures below. The…
Gustav-Gans
  • 365
  • 1
  • 3
  • 13
0
votes
0 answers

Should I always do train/test split before the feature selection process?

I have seen Should Feature Selection be done before Train-Test Split or after? thread and read it. A person had explained there very good. However, is it a must? I mean, if I use the whole dataset for any issue or data without splitting it for…
hotan
  • 19
  • 6
0
votes
1 answer

If the validation isnt converging, my model is overfitting right?

I have this graph, and my validation loss isn't converging. am I right that this is a case of overfitting? graph picture
Sssrnty
  • 3
  • 3
1 2 3
…
9 10