Searched refs:train_loss_tolerance_coeff (Results 1 – 1 of 1) sorted by relevance
172 learning_rate=0.1, rtol=0.2, atol=0.1, train_loss_tolerance_coeff=0.99, argument241 test_case.assertLess(trained_loss * train_loss_tolerance_coeff,244 test_case.assertLess(trained_loss / train_loss_tolerance_coeff,