This is the homework 7 of Introduction to Computer 2021 spring by Prof. Ming-Syan Chen.

In this homework, you are required to build a convolutional neural network for image classification, possibly with some advanced training tips.
Please notice that utilizing external data or pre-trained model for training is prohibited.

### Task Description
Image Classification: Street View House Numbers
Training data: 68257 images
Validation data: 5000 images
Testing data: 26032 images
Extra data: 531131 images

Please **DO NOT** use the testing data and other data, we will consider this as **CHEATING**.

### Evaluation metric: 
Categorization Accuracy

### Deadlines
Kaggle: 6/29 (Tue.) 07:59 in UTC+8 (No late submission!)
Ceiba: 6/29 (Tue.) 23:59 in UTC+8 (Late Submission Policy: 30% off per day)

### Notes
Team Name: **student ID\_anything**, e.g. b09901000\_computer
Maximum Daily Submission: 6

### Contact TAs
Email: computer2021spring@gmail.com
If there is any question or concern, you can discuss on FB group or email us.

### References:
Yuval Netzer, Tao Wang, Adam Coates, Alessandro Bissacco, Bo Wu, Andrew Y. Ng Reading Digits in Natural Images with Unsupervised Feature Learning NIPS Workshop on Deep Learning and Unsupervised Feature Learning 2011.
Link: http://ufldl.stanford.edu/housenumbers