README.md

robot_lab

Please open Problem_1.ipynb on jupyter.

Problem 1 part 1 is finished. You can check the implementation.

SGDM optimizer is implemented.

1. Please load the data first. The you can run the model with SGDM optimizer.
2. After the training, please run the cell below the SGDM cell to save the parameters of the trained model to files.
3. Please run the makefile to compile the c implementation under c_file/ and execute the executable.
4. The loss produced by the trained model implemented in Python should be the same as the one produced by the c implementation.
5. If you want to change the data to test the program, you can change the data set in the python implementation and, the file path and the data set size in the c implementaion in function load_data(). The data set must match.