Todor Davchev

Todor Davchev

PhD student in Robot Learning, University of Edinburgh

Posts by Tags

Deep Learning

Crowded Scene Training/Inference and Useful Tricks

9 minute read

Published:

Now that we have defined the entire model, we can start training the neural network. The idea of each step is to take one batch and predict the next position for each of the agents and positions. The result is then compared to the target values through the associated loss function we defined earlier.

Most Basic Stochastic LSTM for Trajectory Prediction

6 minute read

Published:

In this blog’s experiments we will utilise the mentioned in previous posts (x,y) coordinate representations as input to the network. Since each of these coordinate representations is associated with a specific agent who will interact with each other, it is important to separate the associated sequences and acknowledge that each prediction will be dependent on the previous sequences observed for a given agent.

LSTM

Crowded Scene Training/Inference and Useful Tricks

9 minute read

Published:

Now that we have defined the entire model, we can start training the neural network. The idea of each step is to take one batch and predict the next position for each of the agents and positions. The result is then compared to the target values through the associated loss function we defined earlier.

Most Basic Stochastic LSTM for Trajectory Prediction

6 minute read

Published:

In this blog’s experiments we will utilise the mentioned in previous posts (x,y) coordinate representations as input to the network. Since each of these coordinate representations is associated with a specific agent who will interact with each other, it is important to separate the associated sequences and acknowledge that each prediction will be dependent on the previous sequences observed for a given agent.

Python

Crowded Scene Training/Inference and Useful Tricks

9 minute read

Published:

Now that we have defined the entire model, we can start training the neural network. The idea of each step is to take one batch and predict the next position for each of the agents and positions. The result is then compared to the target values through the associated loss function we defined earlier.

Most Basic Stochastic LSTM for Trajectory Prediction

6 minute read

Published:

In this blog’s experiments we will utilise the mentioned in previous posts (x,y) coordinate representations as input to the network. Since each of these coordinate representations is associated with a specific agent who will interact with each other, it is important to separate the associated sequences and acknowledge that each prediction will be dependent on the previous sequences observed for a given agent.

Stochastic trajectory prediction

Crowded Scene Training/Inference and Useful Tricks

9 minute read

Published:

Now that we have defined the entire model, we can start training the neural network. The idea of each step is to take one batch and predict the next position for each of the agents and positions. The result is then compared to the target values through the associated loss function we defined earlier.

Most Basic Stochastic LSTM for Trajectory Prediction

6 minute read

Published:

In this blog’s experiments we will utilise the mentioned in previous posts (x,y) coordinate representations as input to the network. Since each of these coordinate representations is associated with a specific agent who will interact with each other, it is important to separate the associated sequences and acknowledge that each prediction will be dependent on the previous sequences observed for a given agent.

Tensorflow

Crowded Scene Training/Inference and Useful Tricks

9 minute read

Published:

Now that we have defined the entire model, we can start training the neural network. The idea of each step is to take one batch and predict the next position for each of the agents and positions. The result is then compared to the target values through the associated loss function we defined earlier.

Most Basic Stochastic LSTM for Trajectory Prediction

6 minute read

Published:

In this blog’s experiments we will utilise the mentioned in previous posts (x,y) coordinate representations as input to the network. Since each of these coordinate representations is associated with a specific agent who will interact with each other, it is important to separate the associated sequences and acknowledge that each prediction will be dependent on the previous sequences observed for a given agent.

computer vision

Processing Trajectory Data for Sequence Generation

8 minute read

Published:

Before considering the details around modelling such tasks, we should spend some time to consider the datasets we will use as well as the preprocessing routines we will consider.

data processing

Processing Trajectory Data for Sequence Generation

8 minute read

Published:

Before considering the details around modelling such tasks, we should spend some time to consider the datasets we will use as well as the preprocessing routines we will consider.

stochastic trajectory prediction

Processing Trajectory Data for Sequence Generation

8 minute read

Published:

Before considering the details around modelling such tasks, we should spend some time to consider the datasets we will use as well as the preprocessing routines we will consider.

Tutorial on Stochastic Trajectory Prediction

4 minute read

Published:

This tutorial is meant as an introduction to the problem of trajectory generation. It introduces several ways for modelling the motion of agents in pixel space and proposes several ways of preprocessing data. It follows the structure from its associated GitHub repository. Feel free to skip to the end to see the performance of 2 basic models.

tensorflow

Tutorial on Stochastic Trajectory Prediction

4 minute read

Published:

This tutorial is meant as an introduction to the problem of trajectory generation. It introduces several ways for modelling the motion of agents in pixel space and proposes several ways of preprocessing data. It follows the structure from its associated GitHub repository. Feel free to skip to the end to see the performance of 2 basic models.

trajectory generation

Tutorial on Stochastic Trajectory Prediction

4 minute read

Published:

This tutorial is meant as an introduction to the problem of trajectory generation. It introduces several ways for modelling the motion of agents in pixel space and proposes several ways of preprocessing data. It follows the structure from its associated GitHub repository. Feel free to skip to the end to see the performance of 2 basic models.