convert into the 2d space implement
- [done]run with inital implement
- [done]find the regret net data
- [done]comment the regret net module
- [done]generate our 2d data
- [done]modify the train and test loop
- [pick]use checkpoint to plot the heatmap
- [pick]record the total stats and plot
- [pick]write the experiment part according to the regretnet nm
first get the totally full with the code and the
Dependencies are Python 3, a recent PyTorch, numpy/scipy, tqdm, future and tensorboard.
Plotting with Matplotlib.
Implementation of the neural network is in the module regretnet
.
To train and test auction networks, use the scripts train.py
and test.py
; sample_scripts.sh
gives some examples of how to invoke these.
Pretrained models are in the model
directory.