Course Project - Zero to GANS
##Course Project - Train a Deep Learning Model from Scratch
In this project an attempt made to apply all the concepts learned in this course to train deep learning models end-to-end with PyTorch, and further experimenting with different hyperparameters & metrics to check the effect of selected parameters against model prediction with test data.
The steps followed in this project is outlined as:
- Select a dataset online from https://www.kaggle.com/zalando-research/fashionmnist as per interest.
- Clear description about the data is to be provided.
- Clear description about the type of problem.
- Clear description about the modeling objective.
5.Cleaning of the data is required after checking and to perform exploratory analysis - Creating the “Model” followed by it’s training, and then evaluate against test dataset.
- Saving of all type model data and metrics each time, and repeat the process with different parameters.
- Finally checking of the model performance with different parameters by using “Compare” mode available in Jovian.
!pip install jovian --upgrade --quiet
import jovian
project_name='course_project'
Submission Deadline: Jan 9, 2021 11:59 PM GMT
For the course project, picked a dataset of choice and applied the concepts learned in this course to train deep learning models end-to-end with PyTorch, experimenting with different hyperparameters & metrics.
Example notebooks for reference:
https://jovian.ai/aakashns/simple-cnn-starter
https://jovian.ai/aakashns/transfer-learning-pytorch
https://jovian.ai/aakashns/06b-anime-dcgan
https://jovian.ai/aakashns/05b-cifar10-resnet
I am planning to download the following databases and do project with one:
https://www.kaggle.com/zalando-research/fashionmnist
https://www.kaggle.com/alxmamaev/flowers-recognition
https://www.kaggle.com/paultimothymooney/chest-xray-pneumonia
If all ok then fine.