dmitrySorokin
commented
1 year ago Hi! You can get trained models from this repo (https://bitbucket.org/dmsoroki/interf_game/src/master/). Github doesn't allow to import them due to size limit.
There are three envs in gym_interf repo:
- interf-v1: Mach-Zehnder interferometer without lenses
- interf-v2: Mach-Zehnder interferometer with two lenses (one telescope) in one arm
- interf-v3: Mach-Zehnder interferometer with with two telescopes in one arm
In interf_dqn/discrete_action_wrapper.py it creates "interf-v2" version. Can you explain the error in more details? The code runs fine on my machine.
You can write me at dmitrii.sorokin@phystech.edu
Hi, I am an Imperial College London Physics students working on my master's project and I am really interested in your working and my teammate and I are trying to replicate your simulation. It seems that the trained models you mentioned in the readme files are not included in this version of the repository, we have tried to train it using your dqn training macro but it reported an error on neuron size. Do you ming explaining things in greater details with us if you have time? Your work is really fascinating and we really appreciate if we could learn it better. You can contact me via cf619@ic.ac.uk Best, Chao