https://github.com/dmalexx/binarytrafficsignclassification
BinaryClassification
https://github.com/dmalexx/binarytrafficsignclassification
ai classification keras machine-learning opencv tensorflow
Last synced: 6 months ago
JSON representation
BinaryClassification
- Host: GitHub
- URL: https://github.com/dmalexx/binarytrafficsignclassification
- Owner: DmAlexx
- Created: 2024-03-14T19:23:58.000Z (almost 2 years ago)
- Default Branch: main
- Last Pushed: 2024-03-14T19:43:24.000Z (almost 2 years ago)
- Last Synced: 2025-03-14T03:16:18.195Z (10 months ago)
- Topics: ai, classification, keras, machine-learning, opencv, tensorflow
- Language: Jupyter Notebook
- Homepage:
- Size: 4.1 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
1. Use the data from data/subset_homework folder and visualize some examples. How many images are there for each class?
- class 0 (sign 30) - 2220
- class 1 (sign 50) - 2250
Step 2
Run the training with one single neuron (as we did in the lecture). What accuracy can you achieve?
- Accuracy 0.9044742729306487
Step 3
Make further modifications to improve the accuracy (e.g. add more neurons, more layers, etc.). What is the maximum accuracy you can achieve?
- I've tried some different combination and numbers of neurons in two layers and activation function, but better tnan Accuracy ~ 0.94 I didn't recieve. And it was
- in 2 neurons (linear function) in first dense and 1 neurons (relu function) in second. Other increasing of neurons and layers show me worse results.