Claes Fransson

Thanks a lot for all the explanations and suggestions of modifications to try! :) I have a quite a bit to read/look at and to try to understand now and...

I try to add user options for the kernel length-scales priors lower and upper bounds to have them automatically logged: [https://github.com/Claes1981/chess-tuning-tools/commit/18f64406af2c0dfb3c7bbd198033184265f4be39](https://github.com/Claes1981/chess-tuning-tools/commit/18f64406af2c0dfb3c7bbd198033184265f4be39) But I think I need help with the exact...

Here are some results after setting `lower_bound=0.05` and `upper_bound=0.3` in bask/utils.py: ``` 2021-01-09 20:04:32,518 DEBUG Got the following tuning settings: {'engine1_tc': '15+15', 'engine2_tc': '1+1', 'rounds': 1} 2021-01-09 20:04:32,518 DEBUG Acquisition...

> Thank you for the experiments, I think we can definitely try to optimize the default parameters. Thanks, do you suggest some different values or some more parameters than the...

Thanks, I will try inserting those values in bask/utils.py in my virtual environment. I assume it is the "d_low" and "d_high" values that I should put as `lower_bound` and `upper_bound`...

Thanks, then I think I had the right settings overnight. Here are now some iterations with ``` roundflat = make_roundflat( lower_bound=0.05, upper_bound=0.5, lower_steepness=2.0, upper_steepness=1.0, ) ``` and nu=1.5: ``` 2021-01-11...

Thanks for investigating the observations! Is it correct to conclude from your results and final plot, that my results right after resuming then probably are more correct, than what I...

Thanks, but since I got longer y-scales at resume compared to the continuous iterations before, even if I set gp-initial-burnin and gp-burnin to the same values, I wonder if the...

> > There should be no difference between the resume process and the main optimization loop. The main difference is the number of iterations. "the number of iterations" = what...

sorry, pressed wrong button...