Open b-kartal opened 5 years ago
Thanks for your question. I have added trained models for all scenarios in saved_models. Note these models were trained with ViZDoom version 1.1.4 (some textures changed in the more recent versions)
Regarding TSNE, in visualization there are two scripts, one to generate and save the PCA parameters, one to create the visualization. You might have to hack bit at the second script to get it working, but it should give you an idea.
Let me know if you have any difficulties.
Awesome! I will give it a try to reproduce and let you know.
Thanks!
I just realized I provided code for analyzing the attention distribution, not the TSNE.
I added the TSNE code here again you will need to modify a bit to get it working, but it should get you on the right track.
Sorry about that, let me know if you have further questions.
Thanks I will test it soon.
The fast environment and tasks seem very promising for fast Deep RL experiments. Can you please also release T-SNE visualisation script along with a few trained models? Thanks!