Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Nov 9, 2020

• Update to TensorFlow 2

  dmorrill10

  committed
  44dd826

  Copy full SHA for 44dd826

• Set style

  dmorrill10

  committed
  80550b0

  Copy full SHA for 80550b0

Commits on Nov 14, 2018

• Convert policy to a TensorFlow tensor in dual_state_value_policy_evaluation_op.

  dmorrill10

  committed
  90a9a50

  Copy full SHA for 90a9a50

• Improved primal methods.

  dmorrill10

  committed
  9fd5f55

  Copy full SHA for 9fd5f55

Commits on Nov 1, 2018

• Allow dual_action_value_policy_evaluation_op to be used with state-action dependent discounting.

  dmorrill10

  committed
  ff548b5

  Copy full SHA for ff548b5

Commits on Oct 28, 2018

• Performance refactoring.

  dmorrill10

  committed
  9cb6f46

  Copy full SHA for 9cb6f46

• Add tensor conversions.

  dmorrill10

  committed
  d46a82e

  Copy full SHA for d46a82e

• Performance refactoring.

  dmorrill10

  committed
  d14aea5

  Copy full SHA for d14aea5

• Make dual_state_value_policy_evaluation_op more efficient.

  dmorrill10

  committed
  93ed659

  Copy full SHA for 93ed659

• Extract batch dimension standardization to its own function.

  dmorrill10

  committed
  92868e5

  Copy full SHA for 92868e5

• Make kofn_action_values more efficient.

  dmorrill10

  committed
  0bed344

  Copy full SHA for 0bed344

• Remove unnecessary reshape.

  dmorrill10

  committed
  9114ad9

  Copy full SHA for 9114ad9

• Make dual_state_value_policy_evaluation_op more efficient.

  dmorrill10

  committed
  ab1987c

  Copy full SHA for ab1987c

• Remove unnecessary transpose.

  dmorrill10

  committed
  163c051

  Copy full SHA for 163c051

• Make dual_action_value_policy_evaluation_op more efficient.

  dmorrill10

  committed
  07fd7de

  Copy full SHA for 07fd7de

Commits on Oct 27, 2018

• Fix argument.

  dmorrill10

  committed
  2ae55d4

  Copy full SHA for 2ae55d4

• Fix typo.

  dmorrill10

  committed
  804fc5d

  Copy full SHA for 804fc5d

• Fix argument.

  dmorrill10

  committed
  7725350

  Copy full SHA for 7725350

• Improve dual value functions, update uncertain_reward_discounted_continuing_kofn.py and add tests.

  dmorrill10

  committed
  e096a4a

  Copy full SHA for e096a4a

• Update UncertainRewardDiscountedContinuingKofnGame.

  dmorrill10

  committed
  f6d9a5a

  Copy full SHA for f6d9a5a

• Speed up dual policy evaluation functions.

  dmorrill10

  committed
  5eb33a5

  Copy full SHA for 5eb33a5

Commits on Oct 19, 2018

• Move some utility functions into tf-contextual_prediction_with_expert_advice.

  dmorrill10

  committed
  3165358

  Copy full SHA for 3165358

• Improve state_distribution and improve the numerical stability of successor rep functions.

  dmorrill10

  committed
  0b369d7

  Copy full SHA for 0b369d7

Commits on Oct 18, 2018

• Compare the L1 norm of the state values instead of action values with the value threshold in generalized_policy_iteration_op.

  dmorrill10

  committed
  d24c8a9

  Copy full SHA for d24c8a9

• Allow a value threshold to be set for generalized_policy_iteration_op.

  dmorrill10

  committed
  8285a1c

  Copy full SHA for 8285a1c

• Fix normalized in graph mode.

  dmorrill10

  committed
  6713a38

  Copy full SHA for 6713a38

• Simplify normalized and refactor a little.

  dmorrill10

  committed
  19d4e00

  Copy full SHA for 19d4e00

• Update primal_action_value_policy_evaluation_op and generalized_policy_iteration_op to use a more convenient transitions and reward format.

  Show description for 9ddd025
  dmorrill10

  committed
  9ddd025

  Copy full SHA for 9ddd025

Commits on Oct 17, 2018

• Allow dual value functions to compute average reward when discount is 1.

  dmorrill10

  committed
  a368fea

  Copy full SHA for a368fea

• Fix UncertainRewardDiscountedContinuingKofnEvEnv#__call__.

  dmorrill10

  committed
  4d4b055

  Copy full SHA for 4d4b055

• Allow an initial M_0 matrix to be specified with state_successor_policy_evaluation_op.

  dmorrill10

  committed
  aa77e89

  Copy full SHA for aa77e89

• Refactor k-of-n games and add an environment to compute EV only in an MDP.

  dmorrill10

  committed
  66b1182

  Copy full SHA for 66b1182

• Add functions and a class to refactor k-of-n procedures.

  dmorrill10

  committed
  bb89095

  Copy full SHA for bb89095

• Add cache decorator.

  dmorrill10

  committed
  de36ce8

  Copy full SHA for de36ce8

Commits on Oct 16, 2018

• Undo mistake in state_action_successor_policy_evaluation_op.

  dmorrill10

  committed
  a2f82a3

  Copy full SHA for a2f82a3

Pagination

Previous

Next