Scribd
Upload
0 views2 pages

Introduction To LLMs

The document outlines Homework 1 for an Introduction to Large Language Models course, with a deadline of November 22, 2023. It consists of three main tasks: training an artificial neural network with given parameters, creating a convolutional neural network for ship classification using a YouTube comment dataset, and further training a model on an automobile dataset. Students are required to follow Python's PEP 8 style guide and object-oriented programming principles while completing the assignments.

Uploaded by

cigsesgi0
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0 views2 pages

Introduction To LLMs

The document outlines Homework 1 for an Introduction to Large Language Models course, with a deadline of November 22, 2023. It consists of three main tasks: training an artificial neural network with given parameters, creating a convolutional neural network for ship classification using a YouTube comment dataset, and further training a model on an automobile dataset. Students are required to follow Python's PEP 8 style guide and object-oriented programming principles while completing the assignments.

Uploaded by

cigsesgi0
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Homework 1 - Introduction to Large Language Models

Deadline: Wednesday, 22.11.2023 23:59

1. Arti cial Neural Networks (4.5 points)


The input and output values for this problem are x1 = 2, x2 = 3, x2 = 7 and
t1 = 0.1, t2 = 0.05.

C= 5 -2 1 0 0

9 -6 D= 2 4 0

6 4 2

I. Initialize weights for the parameters to train (0.5 points)


II. Apply forward propagation through the network to get output values (1
point)
III. De ne the error or cost function and its rst derivatives (MSE) (1 point)
IV. Perform backpropagation through the network to determine error
variants (1 point)
V. Update parameter using error derivative and current value (1 point)
fi
fi
fi
2. Youtube Comment Dataset (2.5 points)
Please download the Ship Classi cation data set from the link:

https://www.kaggle.com/datasets/atifaliak/youtube-comments-dataset/
data

I. Create a convolutional NN algorithm. Train your algorithm for 10 ship


types. (0.5 point)
II. Use an embedding model and apply sentiment classi cation. (1 point)
III. Show your testing accuracy of the classi cation algorithm. (1 point)

Please show all your solutions using screenshots and copy them to your
report.

3. The Ship Dataset Classi cation (3 points)


Please download the Ship Classi cation data set from the link:

https://www.kaggle.com/datasets/oleksandershevchenko/ship-
classi cation-dataset

I. Create a convolutional NN algorithm. Train your algorithm for 10 ship


types. (1 point)
II. Use the pre-trained model in step 1. Then ne-tune for an automobile
dataset (you can collect random images from Google). (1 point)
III. Show your testing accuracy of the classi cation algorithm. (1 point)

Please write your code according to "PEP 8 - Style Guide for Python
Code" and follow the object-oriented programming principles, i.e. do
not use any notebook like Jupyter Notebook. https://www.python.org/
dev/peps/pep-0008/
fi
fi
fi
fi
fi
fi
fi
fi

You might also like