Talking-Tom-Runner

Python Plays Talking Tom Gold Run

Edit

i've successfully trained a very very basic cnn to atleast make moves, but ig the main issue lies around the very low dataset and the simplicity of the CNN
I plan on adding a function where it will take on the prediction of the model on the last 5 frames and then on that basis perform the mean or mode on the model's output and do that action in-game

Name		Name	Last commit message	Last commit date
Latest commit History 19 Commits
Test		Test
models		models
1) collect_data.py		1) collect_data.py
2) balanced_data.py		2) balanced_data.py
3) train_model.py		3) train_model.py
4) test_model.py		4) test_model.py
LICENSE		LICENSE
README.md		README.md
Screen Masking Point.txt		Screen Masking Point.txt
TODO.txt		TODO.txt
Talking_Tom_Gold_Rush.ipynb		Talking_Tom_Gold_Rush.ipynb
Talking_Tom_Gold_Rush_AlexNet.ipynb		Talking_Tom_Gold_Rush_AlexNet.ipynb
balanced_data.py		balanced_data.py
conda env.txt		conda env.txt
create_training_data.py		create_training_data.py
directkeys.py		directkeys.py
draw_lanes.py		draw_lanes.py
environment.yml		environment.yml
getkeys.py		getkeys.py
grabscreen.py		grabscreen.py
image_grab.py		image_grab.py
main.py		main.py
model.py		model.py
requirements.txt		requirements.txt
test_model_for_batch_train.py		test_model_for_batch_train.py