This directory contains results updated in real time from an experiment about the partly supervised learning of representations and concepts.

The code is available on https://fleuret.org/git/culture and a summary of the whole experiment is available as a thread on X.com.

plots.png

accuracy and perplexity plots

w_quizzes/

examples of quizzes from the reference hand-designed tasks.

c_quizzes/

quizzes that have been created by the generative models.

prediction/

prediction on quizzes. Green and red frames indicate what has been predicted and if it is correct.

generation/

generation of quizzes.

other/

other files

20240913-17:27:31 wrote ae_002.pth
20240913-17:27:31 wrote ae_003.pth
20240913-17:27:32 wrote ae_004.pth
20240913-17:27:33 wrote ae_005.pth
20240913-17:27:33 wrote ae_006.pth
20240913-17:27:34 wrote ae_007.pth
20240913-17:27:34 wrote ae_008.pth
20240913-17:27:35 wrote ae_009.pth
20240913-17:27:35 epoch_duration 9min2s next_finish 17:36:37
20240913-17:27:35 wrote state.pth
20240913-17:27:35 --- epoch 583 ----------------------------------------
20240913-17:27:35 current_test_accuracies 0.9477 0.9172 0.9506 0.9460 0.9481 0.9455 0.9536 0.9456 0.9477 0.9328
20240913-17:27:35 time_train=8453 time_c_quizzes=0
20240913-17:27:35 nb_c_quizzes 3293
20240913-17:27:35 dummy_rand tensor([0.0584])
20240913-17:27:35 dummy_rand tensor([0.7006])
20240913-17:29:39 train_loss 583 model 9 0.012872862622883986
20240913-17:29:39 train_loss 583 model 1 0.01143296112013013
20240913-17:29:55 test_loss 583 model 9 0.01035431654817785
20240913-17:29:55 test_loss 583 model 1 0.009346395148463386