Questions tagged [training]

Training is the part of machine learning whereby a model is "trained" on a define portion of a dataset to learn attributes and statistical features of the data. It's counterparts are called Testing and Validation. After training a model is tested and validated on another portion of the dataset.

Filter by
Sorted by
Tagged with
0 votes
1 answer
8 views

How to carry metadata with data examples and labels in python?

Is there a pythonic way to carry around metadata that describes the training examples, such that it preserves (i.e., order) after shuffling and splitting (train/test)? ...
0 votes
1 answer
21 views

How would I train an ai to recognize which gender Snapchats bitmojis have

I'd like to enable my python script to get a vague idea of a bitmojis gender. I found these two gitrepositorrys which are able to differentiate them. How would I use them to have my script recognize ...
1 vote
2 answers
30 views

why it would be improper to compute and use test set means?

I have 2 questions regarding the whole subject of the data set in machine learning and I would be happy to receive an answer :) 1.Why it would be improper to compute and use test set means and ...
  • 11
0 votes
0 answers
6 views

How do I deal with missing values when running/training a model based on past information?

I am trying to train a classification model with one of the features being a vector of the objective function values for the last 8 iterations of a hyper-heuristic search. However, the problem I am ...
0 votes
0 answers
9 views

bad prediction when having noise on the data: LSTM time-series regression

I want to predict the force plate using a smart insole using the LSTM model for time series prediction. the data on the force plate has positive and negative values (I think the resulting positive ...
0 votes
1 answer
13 views

kernel initializers and epochs

In a sample code like this (From Keras library): model.add(Dense(13, input_dim=13, kernel_initializer='normal', activation='relu')) the kernel values will be ...
0 votes
0 answers
21 views

Trying to make a visualization for training performance

I am using scikit learn's BayesianRidge model to fit a regression to tabular data of d features and N sample. I have already tested how well my model performs using a repeated kfold cross validation ...
0 votes
0 answers
3 views

Position of training and evaluation methods

I am new in machine learning and was hoping to get an answer about the following. when training machine learning models, we usually use, the model.train() and model.eval(), so : What kind of change ...
0 votes
0 answers
21 views

Fine tuning Bert both on a large, inaccurate dataset and a small, accurate one

I used BertForTokenClassification from HuggingFace transformers (through the french version CamembertForTokenClassification ) to implement a bibliographic reference parser (a tool to extract the ...
1 vote
1 answer
11 views

Predictive value of short text fields

I am working on a classification model using one of the following three algorithms: RandomForestClassifier, a TensorFlow model and a LogisticRegression model. The data set I am working with has a ...
  • 13
0 votes
0 answers
13 views

Correlation analysis of time-series dataset

Clarifications required to proceed after understanding the correlation analysis for processing the time-series dataset: Does the correlation analysis for time-series data need to be different than it ...
0 votes
0 answers
13 views

While predicting on test dataset using RandomForestClassifier, I get error, "Input contains NaN, infinity or a value too large for dtype('float32')."

After seeing many recommendations on StackOverflow and also here, I did the train test split first and then imputation. ...
0 votes
1 answer
24 views

How does sklearn random forest use features in the form of 1D/2D array instead of a single value during splitting at a node?

For the training of random forest model in sklearn, I understand that for features of a single value, a threshold for splitting the data is determined by minimizing the Gini impurity or maximizing ...
1 vote
1 answer
41 views

What does that mean if the loss looks like this?

I have a problem. I have trained a model. And as you can see, there is a zigzag in the loss. In addition, the validation loss is increasing. What does this mean if you only look at the training curve? ...
  • 89
2 votes
0 answers
35 views

In Orange Data Mining, how do I use results from clustering a training-set to test and score a test-set?

I am performing analysis on the well-known 'Adult' data-set, available on UCI using Orange Data Mining. In a PhD thesis, Pelleg (2004; pg 79) uses unsupervised clustering of the prescribed training ...
0 votes
1 answer
17 views

Why should you decide a re training of a machine learning model by a conversion rate or KPI and not randomly or according to perodic time?

Why is it more beneficial to automatically train a machine model by a conversion rate or by a KPI instead of by periodic time? The following image shows an example of the conversion rate. However, as ...
  • 89
0 votes
0 answers
10 views

How to interpret training and validation loss of DeepAR?

Please bear with me. Its a long but complete post. My questions are: Why does the training loss start to osccilate wildly after some epochs? It is because it has jumped out of a local minima? I tried ...
0 votes
1 answer
22 views

Developing Modified KNN Approach

I want to divide the training set into n partitions further besides testing set. How can I do that? Furthermore, I'm creating these groups in the training set. How ...
  • 1
1 vote
1 answer
30 views

Can we train of a binary classifier with "A" to classify "a"?

I have a maybe naive question about the appropriateness of using binary classifications. This is a hypothetical example, so forgive me if it is too coarse. Let's say I want to train a support vector ...
  • 11
0 votes
0 answers
9 views

Sampling from the training set vs passing through the training set one batch at a time

I have seen model implementations where the authors randomly sample a mini-batch from the training set instead of passing through the whole training set each epoch, one mini-batch at a time. When is ...
0 votes
1 answer
16 views

Is it ok to separate data to train on different time instead of putting all in one go?

So let say i have 10,000 images ready to be trained on. But my GPU cannot handle all of that. So the questions is: Can i train the model 10 times with 1000 images each time, with same epochs and ...
0 votes
0 answers
14 views

Different train test split affects model accuracy

If changing the train test split affects the test model accuracy by 10%, does it indicate that the model is not suitable/ overfitting? If yes, what can I do to overcome it. I have utilized all ...
0 votes
0 answers
39 views

Can a fine-tuning dataset for an LLM be autogenerated in principle?

I want to fine-tune an LLM (like GPT-2) to produce a finite automaton from a textual description. My biggest problem is the lack of a proper training set that maps automatons to descriptions. I was ...
  • 113
2 votes
0 answers
21 views

Discussion about modern deep learning training strategies

Previously I have put a lot of effort into training networks appropriately. However, talking to colleagues, a lot of the things I did may be redundant due to novel optimizers and the theory of deep ...
1 vote
1 answer
75 views

Bertopic with embedding: unable to use find_topic

I've used BERTopic with success for the following tasks: get topics, visualise (topics, barcharts, documents ...) and DTM (extended to get area plot with considerable success). However, I am unable to ...
0 votes
0 answers
14 views

What do results like these imply in a LSTM classification problem?

I am training a LSTM network to learn from multiple time series, and the output from the network should be binary (or equivalently a probability score between [0, 1]...
  • 1
0 votes
1 answer
23 views

RNN for continuous, real-time learning without pre-training

I am learning ML and I'm trying to solve this problem Create a rock paper scissors game where the AI is able to beat the player more than 50% of the time. My initial intuition was to use an RNN with ...
0 votes
1 answer
18 views

Where should I stop training if I want to bag models

Let's say I have a clear case of overfitting where my loss curves look like this (x axis are iterations): Now I would like to try bagging to reduce the variance, where should I stop models training? ...
  • 111
0 votes
1 answer
28 views

How to train a regressor model on data that has duplicate subjects but different records for each?

I am working on a dataset that is as follows (just an example): prop_subj prop_comp bed_subj bath_comp sqft_subj sqft_comp A B 2 1 1002 1006 A C 2 2 1002 1075 A D 2 2 1002 1000 B G 2 1 1002 978 ...
1 vote
1 answer
29 views

Is backpropagation applied every layer the same?

For example, I have layers that are pretrained. But while predicted, the loss is very high. But not because of pre-trained layers. Because of not pretrained layers. Will every layer be affected by ...
  • 101
1 vote
1 answer
57 views

Avoid leakage in NLP extraction

What is best practice for applying traditional NLP extraction techniques a pre-processing for ML models? Given a pipeline: Collect raw data. Parse full data set with a variety of traditional NLP ...
  • 435
0 votes
0 answers
30 views

Different results between training and evaluation phase on the same data

I have trained a CNN and in the training phase I obtained an accuracy of 36.5%. If I call model.predict() on the same test data of the training phase I only obtain ...
0 votes
1 answer
44 views

Weird consequence of not freezing layers in Neural Network

I was researching about "why are we freezing layers" and I came across the answer says "to not lose the information of pre-trained model" But; we are just freezing early layers (I ...
  • 101
0 votes
0 answers
7 views

post img labeling transformations

I am using labelImg.py (https://github.com/heartexlabs/labelImg) to label my training data set for a CNN I have (YOLOv4). in order to save on time, I would like to label all 400(ish) images, then ...
0 votes
0 answers
13 views

Do recommender systems predict only from test sets?

I have read about recommender systems. Something I can't understand is if they predict from the test dataset or if they can predict the training dataset too. I thought they fit the model on the ...
  • 3
0 votes
0 answers
58 views

How ca I reshape y_train , y_validation from train_generator?

I retrained ResNet-50 for iris flower classification in tensorflow using the following code: ...
  • 125
0 votes
1 answer
20 views

How to proceed when training data change frequently (in production)?

I'm working with a Recommendation System that would take as parameters a bunch of "tweets" a user see during his navegation on a mobile app. Every tweet has a property, like a category (...
0 votes
0 answers
54 views

CUDA OOM Error, Memory Allocation Keeps Increasing Every Epoch

Hello I have been trying to get this model to run on a computer vision task and keep getting the usual out-of-memory error. The GPU memory always fills up at the 6th epoch no matter the batch_size ...
0 votes
1 answer
17 views

Distribution / number of examples

Imagine you want to predict if a picture is showing a cat or not. First you train your ML algorithm with examples of pictures of cats and dogs and it works. But then you want to train it to also ...
0 votes
2 answers
54 views

Why is my DCGAN not converging?

I'm training a tf DCGAN on the MVTec hazelnut dataset and I found some difficulties. The problem is that after a lot of epochs the generate does not produce some quality images. My model is the ...
  • 1
0 votes
2 answers
251 views

what does shuffle and seed parameter in Keras image_gen.flow_from_directory() signify?

...
0 votes
0 answers
16 views

Training with code scripts to achieve a specified goal?

I am quite new to machine learning and therefore need to ask if some ideas might be possible. Imagine an application that is managing the state of an Actor by ...
  • 101
0 votes
0 answers
31 views

What does the learning curve indicate?

I am training a deep learning model for traffic prediction. When I use 10 months of data for training (validation split: 10%) and 2 months for testing. The loss curve looks like this: . and the ...
0 votes
0 answers
6 views

Bending Training Loss, what could be the cause?

Hello, during training of one of my models, I observe the following training (blue) and test (orange) loss patterns. At first, the training loss increases, then bends and starts decreasing. Just ...
2 votes
0 answers
121 views

Your input ran out of data; interrupting training

I am trying to train ResNET50 for dogs and cats classification (Tensorflow2.3) using the following code: ...
  • 125
1 vote
1 answer
27 views

Getting equal distributions of data from different input sets

I am new to ML. I am trying to create a training dataset that is equally distributed between multiple lists, each of which have a different kind of data. How can I do this? I looked into ...
0 votes
0 answers
32 views

What should be the correct way to re-train images using Yolov3 darknet?

I have been scrolling the internet to read from the experts about custom training using Yolov3 via Darknet. There were great examples, but now I am at confusion on to perform how a retraining. So ...
0 votes
0 answers
16 views

Keras Tensorflow - CNN training performs well on gaming workstation (Windows) but not on high performance Nvidia DGX (Ubuntu)

I've been training on a local machine with Windows 11 (Version 10.0.22) with a 3070 Ti and recently have been able to access a DGX (Ubuntu 20.04.3) with 4 x V100s. Despite numerous re-runs the model ...
  • 1
0 votes
0 answers
22 views

Partial fit with tslearn clustering methods

Is there a way to use partial_fit with tslearn clustering methods like TimeSeriesKMeans? I ...
0 votes
0 answers
100 views

Training a model with ArcFace layer according to code by 4uiiurz1 compatible with TensorFlow Keras

I am trying to train a model with the ArcFace code taken from https://github.com/4uiiurz1/keras-arcface in which I took the ArcFace layer and added it to the model. I created a small dataset of 4 ...

1
2 3 4 5
13