|
|
0c3c709155
|
Switched submodule to relative url so push mirror works
|
2025-07-12 21:27:08 +00:00 |
|
|
|
5dc1f8c554
|
updated training files submodule with a readme
|
2025-07-12 21:17:00 +00:00 |
|
|
|
30e16213fe
|
Added training_files submodule
|
2025-05-13 02:32:22 +00:00 |
|
|
|
d7422cd99b
|
Moved training files to separate folder in preparation of making training_files a submodule
|
2025-05-13 01:40:59 +00:00 |
|
|
|
1f00ca4da4
|
Inverted pendulum report added. Should be finished
|
2025-04-06 22:23:15 +00:00 |
|
|
|
89c06c5c42
|
Fixed a huge issue where base loss centroid convergence was not using 'one' as its reference loss function, resulting in opposite trends.
|
2025-04-06 20:52:50 +00:00 |
|
|
|
8155b0f7ae
|
Basic report added. Report missing significant number of graphs and conclusions
|
2025-04-01 18:57:36 +00:00 |
|
|
|
aa533f2a7e
|
Changing linear regression back to log based
|
2025-04-01 00:16:31 +00:00 |
|
|
|
a1c931480f
|
Changed linear regression to not account for log axes
|
2025-04-01 00:12:36 +00:00 |
|
|
|
867553353b
|
Generated convergence plots for 'centroid' parameters such as t_median, t_mean, R (later/early).
|
2025-03-31 21:50:30 +00:00 |
|
|
|
4d90689a60
|
Generated plots for time_weighting, time_weighting_learning_rate_sweep, and base_loss_learning_rate_sweep
|
2025-03-30 22:22:10 +00:00 |
|
|
|
bfa0f3fb02
|
Trained additional base loss functions for lr sweep. Generated data for time weighting. Generated composite epoch evolution plots
|
2025-03-29 19:23:52 +00:00 |
|
|
|
e238bed91e
|
Finding best learning rates from the sweeps
|
2025-03-29 02:07:34 +00:00 |
|
|
|
aa34bfac8c
|
Base loss learning rate sweep training done. Made composite plots for epoch evolution of different time weightings
|
2025-03-26 04:20:07 +00:00 |
|
|
|
a401ca3f59
|
Done re-training max_normalized time weighting (with a minimum weight of 0.01) and time_weight_learning_sweep). Started work on base loss function training
|
2025-03-12 22:33:37 +00:00 |
|
|
|
eb71ab0de9
|
Finish training with mirrored weights. Plot max normalized with mirror weights
|
2025-02-25 23:27:49 +00:00 |
|
|
|
68f918f51c
|
Added mirror weight functions
|
2025-02-25 04:20:15 +00:00 |
|
|
|
c89998da28
|
Fix issues where the initialized controllers could be different. Created a controller_base.pth that is used for all controller initialization
|
2025-02-24 02:28:16 +00:00 |
|
|
|
8be7ad97a8
|
Plotting theta across epochs for the different loss functions
|
2025-02-22 23:45:53 +00:00 |
|
|
|
7d4d34a580
|
Better analysis file structure
|
2025-02-19 04:02:53 +00:00 |
|
|
|
b9212e5a52
|
Average normalized results across epochs have been plotted
|
2025-02-19 03:06:48 +00:00 |
|
|
|
8f92ce3ee1
|
Restructure files. Changed weight functions to be normalized from 0-1ns to always be normalized from 0-1 (aka max normalization). Also updated average normalization
|
2025-02-18 18:45:15 +00:00 |
|
|
|
28c5d14fe8
|
Plotted controller max normalzied across epoch. Also training average normalized
|
2025-02-18 00:40:29 +00:00 |
|
|
|
071669696b
|
Completed training for max normalized comparison
|
2025-02-17 18:47:05 +00:00 |
|
|
|
3614b66aee
|
Normalization of loss functions based on max weight
|
2025-02-17 02:54:56 +00:00 |
|
|
|
3fc78d4508
|
Looking at the different controllers as they trained
|
2025-02-16 00:03:29 +00:00 |
|
|
|
a865d37722
|
Controller across epochs plotter
|
2025-02-15 23:07:51 +00:00 |
|
|
|
dd97ce7335
|
Inverted pendulum controllers trained
|
2025-02-15 16:16:09 +00:00 |
|
|
|
ceda15213b
|
Trained inverted pendulum, no time weights
|
2025-02-06 13:39:53 +00:00 |
|
|
|
0bb316f052
|
Redo training files to save models after every epoch
|
2025-02-06 03:10:29 +00:00 |
|
|
|
5f70241418
|
Inverted pendulum training started for no time weight, linear, quadratic, cubic, and exponential
|
2025-02-05 05:07:15 +00:00 |
|
|
|
a6273835b1
|
Inverted pendulum with desired theta trained
|
2025-02-04 03:57:47 +00:00 |
|
|
|
cdefc00226
|
Adding 'desired_theta' as neural network input
|
2025-02-03 22:28:41 +00:00 |
|
|
|
1b7b40adbc
|
Inverse pendulum testing
|
2025-02-01 20:34:31 +00:00 |
|
|
|
f76fa8709d
|
Inverse pendulum testing
|
2025-02-01 20:34:19 +00:00 |
|
