author int64 658 755k | date stringdate 2012-06-12 08:34:29 2024-07-22 14:51:21 | timezone int64 -46,800 43.2k | hash stringlengths 40 40 | message stringlengths 5 490 | mods listlengths 1 16 | language stringclasses 20
values | license stringclasses 3
values | repo stringlengths 5 68 | original_message stringlengths 12 491 |
|---|---|---|---|---|---|---|---|---|---|
641,014 | 08.07.2021 13:33:01 | -32,400 | 526c159c8c8e8f7a2a4e7647984ca5631a66d557 | fix tests of inits and add some descriptions | [
{
"change_type": "MODIFY",
"old_path": "tests/ope/test_dr_estimators.py",
"new_path": "tests/ope/test_dr_estimators.py",
"diff": "@@ -14,6 +14,82 @@ from obp.ope import (\n)\nfrom conftest import generate_action_dist\n+\n+invalid_input_of_dr_init = [\n+ (\n+ \"\",\n+ TypeError,\n+ r\"`lambda_` must ... | Python | Apache License 2.0 | st-tech/zr-obp | fix tests of inits and add some descriptions |
641,014 | 08.07.2021 19:45:12 | -32,400 | cbc33686e8cf2c60144cce67ab7452feb4b146b2 | add checks wrt np.nan | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/estimators.py",
"new_path": "obp/ope/estimators.py",
"diff": "@@ -258,7 +258,7 @@ class InverseProbabilityWeighting(BaseOffPolicyEstimator):\nwhere :math:`\\\\mathcal{D}=\\\\{(x_t,a_t,r_t)\\\\}_{t=1}^{T}` is logged bandit feedback data with :math:`T` r... | Python | Apache License 2.0 | st-tech/zr-obp | add checks wrt np.nan |
641,014 | 09.07.2021 19:21:05 | -32,400 | ebc94523b642e1c677e35c5946c2380bd835c0b3 | implement _estimate_mse_upper_bound method | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/estimators.py",
"new_path": "obp/ope/estimators.py",
"diff": "@@ -537,6 +537,71 @@ class InverseProbabilityWeighting(BaseOffPolicyEstimator):\nrandom_state=random_state,\n)\n+ def _estimate_mse_upper_bound(\n+ self,\n+ reward: np.ndarray,\n+ action: np... | Python | Apache License 2.0 | st-tech/zr-obp | implement _estimate_mse_upper_bound method |
641,014 | 09.07.2021 19:28:33 | -32,400 | 0a8b7edc3d6ea6e9043bce15bce1fb5493cc92bb | add some tests and rename variables | [
{
"change_type": "MODIFY",
"old_path": "tests/ope/test_dr_estimators.py",
"new_path": "tests/ope/test_dr_estimators.py",
"diff": "@@ -207,9 +207,9 @@ def test_dr_tuning_init_using_valid_input_data(lambdas_taus, description):\ndm = DirectMethod()\ndr = DoublyRobust()\ndr_tuning = DoublyRobustTuning(l... | Python | Apache License 2.0 | st-tech/zr-obp | add some tests and rename variables |
641,014 | 10.07.2021 10:25:24 | -32,400 | 089c6f16140998c8bffeba556fbfddcf294316dd | use dict to store mse upper bounds | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/estimators_tuning.py",
"new_path": "obp/ope/estimators_tuning.py",
"diff": "@@ -125,7 +125,7 @@ class InverseProbabilityWeightingTuning(InverseProbabilityWeighting):\nposition = np.zeros(action_dist.shape[0], dtype=int)\n# tune the clipping hyperparame... | Python | Apache License 2.0 | st-tech/zr-obp | use dict to store mse upper bounds |
641,014 | 10.07.2021 14:55:10 | -32,400 | f1a1259686c6c00b4e535862d020e3c889d9912c | apply direct bias estimation | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/estimators.py",
"new_path": "obp/ope/estimators.py",
"diff": "@@ -568,10 +568,6 @@ class InverseProbabilityWeighting(BaseOffPolicyEstimator):\nposition: array-like, shape (n_rounds,), default=None\nPositions of each round in the given logged bandit fee... | Python | Apache License 2.0 | st-tech/zr-obp | apply direct bias estimation |
641,014 | 10.07.2021 15:49:44 | -32,400 | df8517000ab794f0860147b731d5b197a252d830 | add high prob deviation bound term to estiamte bias upper bound | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/estimators.py",
"new_path": "obp/ope/estimators.py",
"diff": "@@ -593,7 +593,11 @@ class InverseProbabilityWeighting(BaseOffPolicyEstimator):\niw = action_dist[np.arange(n_rounds), action, position] / pscore\niw_hat = np.minimum(iw, self.lambda_)\nbias... | Python | Apache License 2.0 | st-tech/zr-obp | add high prob deviation bound term to estiamte bias upper bound |
641,006 | 11.07.2021 15:42:54 | -32,400 | 01b1fdc1fad16376cd125802e4fdd620a600d390 | rephrase polict_logit -> policy_softmax; fix typo comment | [
{
"change_type": "MODIFY",
"old_path": "obp/dataset/synthetic_slate.py",
"new_path": "obp/dataset/synthetic_slate.py",
"diff": "@@ -359,8 +359,8 @@ class SyntheticSlateBanditDataset(BaseBanditDataset):\nreturn pscores\n- def _calc_pscore_given_policy_value(\n- self, all_slate_actions: np.ndarray, po... | Python | Apache License 2.0 | st-tech/zr-obp | rephrase polict_logit -> policy_softmax; fix typo comment |
641,014 | 12.07.2021 18:54:57 | -32,400 | 3a04cd8eda180028191355a1af65f87720d3aec4 | rm some tests as new classes do not have to adjust to tensor inputs | [
{
"change_type": "MODIFY",
"old_path": "tests/ope/test_dr_estimators.py",
"new_path": "tests/ope/test_dr_estimators.py",
"diff": "@@ -215,13 +215,17 @@ def test_dr_tuning_init_using_valid_input_data(lambdas_taus, description):\n# prepare instances\ndm = DirectMethod()\ndr = DoublyRobust()\n-dr_tunin... | Python | Apache License 2.0 | st-tech/zr-obp | rm some tests as new classes do not have to adjust to tensor inputs |
641,014 | 12.07.2021 19:16:21 | -32,400 | 7acd210bbe2c0e7a76c2073b58deb14c1556e83f | fix a wrong wording | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/estimators.py",
"new_path": "obp/ope/estimators.py",
"diff": "@@ -205,7 +205,7 @@ class ReplayMethod(BaseOffPolicyEstimator):\nPositions of each round in the given logged bandit feedback.\nalpha: float, default=0.05\n- P-value.\n+ Significance level.\n... | Python | Apache License 2.0 | st-tech/zr-obp | fix a wrong wording |
641,014 | 13.07.2021 12:56:51 | -32,400 | 97a569e36c0651a766814e2113bd8f508e8006dc | add estimators with hyperparam tuning | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/__init__.py",
"new_path": "obp/ope/__init__.py",
"diff": "@@ -50,3 +50,11 @@ __all_estimators__ = [\n\"SwitchDoublyRobust\",\n\"SelfNormalizedDoublyRobust\",\n]\n+\n+\n+__all_estimators_tuning__ = [\n+ \"InverseProbabilityWeightingTuning\",\n+ \"Doubly... | Python | Apache License 2.0 | st-tech/zr-obp | add estimators with hyperparam tuning |
641,014 | 18.07.2021 14:41:23 | -32,400 | c6284f176fdd58021ad0e1dbe563188d0eddbfb5 | add subsample option to OpenBanditDataset | [
{
"change_type": "MODIFY",
"old_path": "obp/dataset/real.py",
"new_path": "obp/dataset/real.py",
"diff": "from dataclasses import dataclass\nfrom logging import getLogger, basicConfig, INFO\nfrom pathlib import Path\n-from typing import Optional\n-from typing import Union\n-from typing import Tuple\... | Python | Apache License 2.0 | st-tech/zr-obp | add subsample option to OpenBanditDataset |
641,014 | 18.07.2021 14:41:41 | -32,400 | 806c639b4c2416d75b3a73b157f92c82b230e7e9 | add some tests of OpenBanditDataset | [
{
"change_type": "MODIFY",
"old_path": "tests/dataset/test_real.py",
"new_path": "tests/dataset/test_real.py",
"diff": "@@ -19,9 +19,7 @@ def test_real_init():\n# data_path\nwith pytest.raises(ValueError):\n- OpenBanditDataset(\n- behavior_policy=\"random\", campaign=\"all\", data_path=\"raw_str_pat... | Python | Apache License 2.0 | st-tech/zr-obp | add some tests of OpenBanditDataset |
641,014 | 28.08.2021 12:17:40 | 14,400 | 962a92f9a231065b1c73cf8573f36771babb541d | add some arguments | [
{
"change_type": "MODIFY",
"old_path": "obp/dataset/synthetic.py",
"new_path": "obp/dataset/synthetic.py",
"diff": "@@ -7,11 +7,11 @@ from typing import Optional, Callable\nimport numpy as np\nfrom scipy.stats import truncnorm\n-from sklearn.utils import check_random_state\n+from sklearn.utils impor... | Python | Apache License 2.0 | st-tech/zr-obp | add some arguments |
641,014 | 28.08.2021 12:17:55 | 14,400 | 501b0c9a818cf43af4f30ecebea759e401c49e8e | add sample_action_fact func | [
{
"change_type": "MODIFY",
"old_path": "obp/utils.py",
"new_path": "obp/utils.py",
"diff": "@@ -93,6 +93,33 @@ def estimate_confidence_interval_by_bootstrap(\n}\n+def sample_action_fast(\n+ action_dist: np.ndarray, random_state: Optional[int] = None\n+) -> np.ndarray:\n+ \"\"\"Sample actions faster ... | Python | Apache License 2.0 | st-tech/zr-obp | add sample_action_fact func |
641,014 | 28.08.2021 12:18:11 | 14,400 | d15e992a5cca7ab8ca1a45a3294c4232ab4c7700 | add some tests about SyntheticBanditDataset | [
{
"change_type": "MODIFY",
"old_path": "tests/dataset/test_synthetic.py",
"new_path": "tests/dataset/test_synthetic.py",
"diff": "@@ -28,6 +28,20 @@ def test_synthetic_init():\nwith pytest.raises(ValueError):\nSyntheticBanditDataset(n_actions=2, reward_type=\"aaa\")\n+ # reward_std\n+ with pytest.ra... | Python | Apache License 2.0 | st-tech/zr-obp | add some tests about SyntheticBanditDataset |
641,014 | 28.08.2021 12:53:01 | 14,400 | d4d4684e06c5fed117b383a030e98bfd9f9c75bc | fix a bug and typo in tests | [
{
"change_type": "MODIFY",
"old_path": "obp/dataset/synthetic.py",
"new_path": "obp/dataset/synthetic.py",
"diff": "@@ -465,7 +465,7 @@ def linear_behavior_policy(\nReturns\n---------\nbehavior_policy: array-like, shape (n_rounds, n_actions)\n- Action choice probabilities given context (:math:`x`), ... | Python | Apache License 2.0 | st-tech/zr-obp | fix a bug and typo in tests |
641,014 | 30.08.2021 07:28:19 | 14,400 | 020a78ee3e90c9cc55896220f2d86d62dcbdd883 | fix Error of meta | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/meta.py",
"new_path": "obp/ope/meta.py",
"diff": "@@ -12,7 +12,7 @@ import numpy as np\nfrom pandas import DataFrame\nimport seaborn as sns\n-from .estimators import BaseOffPolicyEstimator\n+from .estimators import BaseOffPolicyEstimator, DirectMethod ... | Python | Apache License 2.0 | st-tech/zr-obp | fix Error of meta |
641,014 | 30.08.2021 14:36:02 | 14,400 | 36beaad6a1c9e6c5c4aee8a572338138645f7fc9 | add check_array to synthetic_continuous | [
{
"change_type": "MODIFY",
"old_path": "obp/dataset/synthetic_continuous.py",
"new_path": "obp/dataset/synthetic_continuous.py",
"diff": "@@ -11,6 +11,7 @@ from sklearn.utils import check_random_state, check_scalar\nfrom .base import BaseBanditDataset\nfrom ..types import BanditFeedback\n+from ..uti... | Python | Apache License 2.0 | st-tech/zr-obp | add check_array to synthetic_continuous |
641,014 | 31.08.2021 18:18:12 | 14,400 | 274f651a48f3f2e088a09b5b9d44e350e1855c1e | add error to detect negative rewards for IPWLearner | [
{
"change_type": "MODIFY",
"old_path": "obp/policy/offline.py",
"new_path": "obp/policy/offline.py",
"diff": "@@ -18,8 +18,7 @@ import torch.optim as optim\nfrom tqdm import tqdm\nfrom .base import BaseOfflinePolicyLearner\n-\n-from ..utils import check_bandit_feedback_inputs\n+from ..utils import c... | Python | Apache License 2.0 | st-tech/zr-obp | add error to detect negative rewards for IPWLearner |
641,014 | 31.08.2021 18:18:28 | 14,400 | e3f494d818a5a2e32f582ee0cfd1da3f7d895b9a | adjust tests to changes in obp | [
{
"change_type": "MODIFY",
"old_path": "tests/ope/test_regression_models.py",
"new_path": "tests/ope/test_regression_models.py",
"diff": "@@ -334,7 +334,7 @@ invalid_input_of_fitting_regression_models = [\nNone,\n3,\n1,\n- \"context, action, reward, and position must have the same number of samples.... | Python | Apache License 2.0 | st-tech/zr-obp | adjust tests to changes in obp |
641,014 | 31.08.2021 18:28:07 | 14,400 | 35a053fed6a0fe1f8155e8a5c6bf0e0604995af0 | apply check_array functions | [
{
"change_type": "MODIFY",
"old_path": "obp/dataset/synthetic_continuous.py",
"new_path": "obp/dataset/synthetic_continuous.py",
"diff": "@@ -240,12 +240,11 @@ class SyntheticContinuousBanditDataset(BaseBanditDataset):\n\"\"\"\ncheck_array(array=context, name=\"context\", expected_dim=2)\n+ check_ar... | Python | Apache License 2.0 | st-tech/zr-obp | apply check_array functions |
641,014 | 03.09.2021 18:16:10 | 14,400 | 67d283af7f16281c7445cce5bd43972df8321e14 | make the hyperparameter tuning method more flexible | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/helper.py",
"new_path": "obp/ope/helper.py",
"diff": "@@ -7,6 +7,53 @@ import numpy as np\nfrom sklearn.utils import check_scalar\n+def estimate_bias_in_ope(\n+ reward: np.ndarray,\n+ iw: np.ndarray,\n+ iw_hat: np.ndarray,\n+ q_hat: Optional[np.ndarray... | Python | Apache License 2.0 | st-tech/zr-obp | make the hyperparameter tuning method more flexible |
641,014 | 03.09.2021 18:45:16 | 14,400 | 2f0e1ed01140fc9724d8efbc62eb7f9501ad5ba2 | change tau of Switch to lambda_ to make the name consistent across estimators | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/estimators.py",
"new_path": "obp/ope/estimators.py",
"diff": "@@ -1521,20 +1521,20 @@ class SwitchDoublyRobust(DoublyRobust):\n.. math::\n- \\\\hat{V}_{\\\\mathrm{SwitchDR}} (\\\\pi_e; \\\\mathcal{D}, \\\\hat{q}, \\\\tau)\n- := \\\\mathbb{E}_{\\\\mathc... | Python | Apache License 2.0 | st-tech/zr-obp | change tau of Switch to lambda_ to make the name consistent across estimators |
641,014 | 03.09.2021 18:56:09 | 14,400 | 175b70d7073a8fa05bc22a3dc1e0221ba9ce2527 | adjust to the change of the name of the argument | [
{
"change_type": "MODIFY",
"old_path": "examples/multiclass/README.md",
"new_path": "examples/multiclass/README.md",
"diff": "@@ -76,8 +76,8 @@ python evaluate_off_policy_estimators.py\\\n# snipw 0.006797 0.004094\n# dr 0.007780 0.004492\n# sndr 0.007210 0.004089\n-# switch-dr (tau=1) 0.173282 0.020... | Python | Apache License 2.0 | st-tech/zr-obp | adjust to the change of the name of the argument |
641,014 | 03.09.2021 20:07:39 | 14,400 | e488b0e8033f29221a203069ad7a26bf6f11c302 | adjust the benchmark code to the change in hyperparam name | [
{
"change_type": "MODIFY",
"old_path": "benchmark/ope/README.md",
"new_path": "benchmark/ope/README.md",
"diff": "@@ -90,7 +90,7 @@ It is possible to run multiple experimental settings easily by using the `--mult\nFor example, the following script sweeps over all simulations including the three camp... | Python | Apache License 2.0 | st-tech/zr-obp | adjust the benchmark code to the change in hyperparam name |
641,014 | 04.09.2021 14:31:53 | 14,400 | 4ed27edf8c09cb562eabe92f209195454355f5c1 | modify defaults and fix bugs | [
{
"change_type": "MODIFY",
"old_path": "obp/policy/offline.py",
"new_path": "obp/policy/offline.py",
"diff": "@@ -417,7 +417,7 @@ class NNPolicyLearner(BaseOfflinePolicyLearner):\nlearning_rate_init: int, default=0.0001\nInitial learning rate for SGD, Adagrad, and Adam.\n- max_iter: int, default=200... | Python | Apache License 2.0 | st-tech/zr-obp | modify defaults and fix bugs |
641,006 | 05.09.2021 15:42:58 | -32,400 | 97cadef8234ffcee3d73365a84c67f6d400264b6 | fix multiclass.py | [
{
"change_type": "MODIFY",
"old_path": "obp/dataset/multiclass.py",
"new_path": "obp/dataset/multiclass.py",
"diff": "@@ -9,7 +9,7 @@ import numpy as np\nfrom scipy.stats import rankdata\nfrom sklearn.base import ClassifierMixin, is_classifier, clone\nfrom sklearn.model_selection import train_test_s... | Python | Apache License 2.0 | st-tech/zr-obp | fix multiclass.py |
641,006 | 05.09.2021 15:43:24 | -32,400 | 1860af79db6d3672b6270658a86796be3bfbf1f4 | fix synthetic.py and the test | [
{
"change_type": "MODIFY",
"old_path": "obp/dataset/synthetic.py",
"new_path": "obp/dataset/synthetic.py",
"diff": "@@ -146,14 +146,8 @@ class SyntheticBanditDataset(BaseBanditDataset):\ndef __post_init__(self) -> None:\n\"\"\"Initialize Class.\"\"\"\n- if not isinstance(self.n_actions, int) or self... | Python | Apache License 2.0 | st-tech/zr-obp | fix synthetic.py and the test |
641,006 | 05.09.2021 16:00:33 | -32,400 | 60da3817848c2b2f18af48a4ecc19b2b347e967c | fix synthetic_slate.py and the test | [
{
"change_type": "MODIFY",
"old_path": "obp/dataset/synthetic_slate.py",
"new_path": "obp/dataset/synthetic_slate.py",
"diff": "@@ -190,24 +190,14 @@ class SyntheticSlateBanditDataset(BaseBanditDataset):\ndef __post_init__(self) -> None:\n\"\"\"Initialize Class.\"\"\"\n- if not isinstance(self.n_uni... | Python | Apache License 2.0 | st-tech/zr-obp | fix synthetic_slate.py and the test |
641,006 | 05.09.2021 17:01:08 | -32,400 | 8578e0c4930fc57edf45d635b2e8cda81bb387a4 | add in error message | [
{
"change_type": "MODIFY",
"old_path": "obp/dataset/synthetic_continuous.py",
"new_path": "obp/dataset/synthetic_continuous.py",
"diff": "@@ -243,7 +243,7 @@ class SyntheticContinuousBanditDataset(BaseBanditDataset):\ncheck_array(array=action, name=\"action\", expected_dim=1)\nif context.shape[1] !=... | Python | Apache License 2.0 | st-tech/zr-obp | add in error message |
641,006 | 05.09.2021 17:51:49 | -32,400 | f8a4832440bcf5a64299833b6e8fbcc7f4d23d67 | fix contextfree check-scalar | [
{
"change_type": "MODIFY",
"old_path": "obp/policy/base.py",
"new_path": "obp/policy/base.py",
"diff": "@@ -7,7 +7,7 @@ from dataclasses import dataclass\nfrom typing import Optional\nimport numpy as np\n-from sklearn.utils import check_random_state\n+from sklearn.utils import check_random_state, ch... | Python | Apache License 2.0 | st-tech/zr-obp | fix contextfree check-scalar |
641,006 | 05.09.2021 17:52:59 | -32,400 | 7d8d0ceaa389876b7c7415318825d947fad269f7 | fix linear and logistic check-scalar | [
{
"change_type": "MODIFY",
"old_path": "obp/policy/linear.py",
"new_path": "obp/policy/linear.py",
"diff": "from dataclasses import dataclass\nimport numpy as np\n+from sklearn.utils import check_scalar\nfrom .base import BaseContextualPolicy\n+from ..utils import check_array\n@dataclass\n@@ -123,10... | Python | Apache License 2.0 | st-tech/zr-obp | fix linear and logistic check-scalar |
641,006 | 05.09.2021 17:53:13 | -32,400 | e979ca8f3d7cd538f77085f63ca88d508d435314 | fix offline check-scalar | [
{
"change_type": "MODIFY",
"old_path": "obp/policy/offline.py",
"new_path": "obp/policy/offline.py",
"diff": "@@ -484,10 +484,7 @@ class NNPolicyLearner(BaseOfflinePolicyLearner):\nif self.len_list != 1:\nraise NotImplementedError(\"currently, len_list > 1 is not supported\")\n- if not isinstance(se... | Python | Apache License 2.0 | st-tech/zr-obp | fix offline check-scalar |
641,006 | 05.09.2021 20:07:57 | -32,400 | 1eeb28fc644ff6160e9406d221b2649d4d5bfcad | fix offline-continuous check-scalar | [
{
"change_type": "MODIFY",
"old_path": "obp/policy/offline_continuous.py",
"new_path": "obp/policy/offline_continuous.py",
"diff": "@@ -8,7 +8,7 @@ from typing import Tuple, Optional, Union, Dict\nfrom tqdm import tqdm\nimport numpy as np\n-from sklearn.utils import check_random_state\n+from sklearn... | Python | Apache License 2.0 | st-tech/zr-obp | fix offline-continuous check-scalar |
641,006 | 06.09.2021 00:46:43 | -32,400 | 0e28a8c2ab696f0712d1725bbc583644556c4771 | add self-normalized slate estimators | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/__init__.py",
"new_path": "obp/ope/__init__.py",
"diff": "@@ -15,6 +15,8 @@ from obp.ope.estimators_slate import SlateStandardIPS\nfrom obp.ope.estimators_slate import SlateIndependentIPS\nfrom obp.ope.estimators_slate import SlateRewardInteractionIPS\... | Python | Apache License 2.0 | st-tech/zr-obp | add self-normalized slate estimators |
641,014 | 06.09.2021 20:21:14 | 14,400 | 16c7d1c20f9ac30bf6327ecd967d7e46c36c48f9 | update obp version in benchmark | [
{
"change_type": "MODIFY",
"old_path": "benchmark/ope/README.md",
"new_path": "benchmark/ope/README.md",
"diff": "@@ -53,8 +53,8 @@ numpy = \"^1.21.1\"\nmatplotlib = \"^3.4.2\"\nhydra-core = \"^1.1.0\"\npingouin = \"^0.4.0\"\n-obp = \"^0.4.2\"\npyieoe = \"^0.1.0\"\n+obp = \"^0.5.0\"\n```\n## Files\n... | Python | Apache License 2.0 | st-tech/zr-obp | update obp version in benchmark |
641,006 | 20.09.2021 17:08:24 | -32,400 | 5deff62cf41ba553b25060dcebdd461734a04675 | fix self normalized estimator class | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/estimators_slate.py",
"new_path": "obp/ope/estimators_slate.py",
"diff": "@@ -128,44 +128,6 @@ class BaseSlateInverseProbabilityWeighting(BaseSlateOffPolicyEstimator):\nrandom_state=random_state,\n)\n- def _estimate_round_rewards_self_normalized(\n- se... | Python | Apache License 2.0 | st-tech/zr-obp | fix self normalized estimator class |
641,006 | 20.09.2021 17:08:43 | -32,400 | 7957c278beddd7b1224c70a0283a617d1f0e3696 | add self normalized slate ope tests | [
{
"change_type": "MODIFY",
"old_path": "tests/ope/test_ipw_estimators_slate.py",
"new_path": "tests/ope/test_ipw_estimators_slate.py",
"diff": "import pytest\nimport numpy as np\n-from obp.ope import SlateStandardIPS, SlateIndependentIPS, SlateRewardInteractionIPS\n+from obp.ope import (\n+ SlateSta... | Python | Apache License 2.0 | st-tech/zr-obp | add self normalized slate ope tests |
641,014 | 27.09.2021 15:35:47 | -7,200 | 5a956b322bf8df7e310a8f627445eabb95c3d2fd | fix incorrect metadata | [
{
"change_type": "MODIFY",
"old_path": "pyproject.toml",
"new_path": "pyproject.toml",
"diff": "[tool.poetry]\nname = \"obp\"\n-version = \"0.4.2\"\n+version = \"0.5.1\"\ndescription = \"Open Bandit Pipeline: a python library for bandit algorithms and off-policy evaluation\"\nauthors = [\"Yuta Saito... | Python | Apache License 2.0 | st-tech/zr-obp | fix incorrect metadata |
641,014 | 26.10.2021 07:56:30 | 14,400 | 240a6ff429c1859e320d29a43aea587a2c5b37ef | add q-learner | [
{
"change_type": "MODIFY",
"old_path": "obp/policy/__init__.py",
"new_path": "obp/policy/__init__.py",
"diff": "@@ -13,6 +13,7 @@ from obp.policy.logistic import LogisticTS\nfrom obp.policy.logistic import LogisticUCB\nfrom obp.policy.logistic import MiniBatchLogisticRegression\nfrom obp.policy.offl... | Python | Apache License 2.0 | st-tech/zr-obp | add q-learner |
641,014 | 26.10.2021 16:43:09 | 14,400 | fdd4ca2814cc91720efaf6190e5753e7a36f61eb | unify some test funcs | [
{
"change_type": "MODIFY",
"old_path": "tests/policy/test_offline.py",
"new_path": "tests/policy/test_offline.py",
"diff": "@@ -5,8 +5,8 @@ from sklearn.linear_model import LogisticRegression\nimport torch\nfrom obp.policy.offline import IPWLearner\n-from obp.policy.offline import QLearner\nfrom obp... | Python | Apache License 2.0 | st-tech/zr-obp | unify some test funcs |
641,014 | 29.10.2021 23:53:31 | 14,400 | d5d107003f11ed37d1e4ff834cee02137e085278 | implement the gumble softmax trick to sample rankings of actions | [
{
"change_type": "MODIFY",
"old_path": "obp/policy/offline.py",
"new_path": "obp/policy/offline.py",
"diff": "@@ -29,7 +29,6 @@ from obp.ope import RegressionModel\nfrom ..utils import check_array\nfrom ..utils import check_bandit_feedback_inputs\nfrom ..utils import check_tensor\n-from ..utils impo... | Python | Apache License 2.0 | st-tech/zr-obp | implement the gumble softmax trick to sample rankings of actions |
641,014 | 10.11.2021 21:04:00 | 18,000 | a67356b2471c86d1f04a3594fabb8b2e65007c1b | modify reward and behavior policy functions | [
{
"change_type": "MODIFY",
"old_path": "obp/dataset/synthetic.py",
"new_path": "obp/dataset/synthetic.py",
"diff": "@@ -8,6 +8,7 @@ from typing import Optional\nimport numpy as np\nfrom scipy.stats import truncnorm\n+from sklearn.preprocessing import PolynomialFeatures\nfrom sklearn.utils import che... | Python | Apache License 2.0 | st-tech/zr-obp | modify reward and behavior policy functions |
641,014 | 12.11.2021 10:02:19 | 18,000 | 99c0d632a72b8168a8823e9674ec51bbe12462f3 | define base reward and behavior policy functions | [
{
"change_type": "MODIFY",
"old_path": "obp/dataset/__init__.py",
"new_path": "obp/dataset/__init__.py",
"diff": "@@ -4,7 +4,10 @@ from obp.dataset.multiclass import MultiClassToBanditReduction\nfrom obp.dataset.real import OpenBanditDataset\nfrom obp.dataset.synthetic import linear_behavior_policy\... | Python | Apache License 2.0 | st-tech/zr-obp | define base reward and behavior policy functions |
641,014 | 13.11.2021 03:30:28 | 18,000 | a499ec2f89046037657b79e78a5fd993edccd66c | add sparse reward functions | [
{
"change_type": "MODIFY",
"old_path": "obp/dataset/__init__.py",
"new_path": "obp/dataset/__init__.py",
"diff": "@@ -6,8 +6,10 @@ from obp.dataset.synthetic import linear_behavior_policy\nfrom obp.dataset.synthetic import linear_reward_function\nfrom obp.dataset.synthetic import logistic_polynomial... | Python | Apache License 2.0 | st-tech/zr-obp | add sparse reward functions |
641,006 | 18.11.2021 17:06:54 | -32,400 | 55653a8b6444774705fa88bf6b889b024e1ea5ce | initial commit of b-ope | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/__init__.py",
"new_path": "obp/ope/__init__.py",
"diff": "@@ -7,6 +7,7 @@ from obp.ope.estimators import ReplayMethod\nfrom obp.ope.estimators import SelfNormalizedDoublyRobust\nfrom obp.ope.estimators import SelfNormalizedInverseProbabilityWeighting\n... | Python | Apache License 2.0 | st-tech/zr-obp | initial commit of b-ope |
641,006 | 26.11.2021 21:09:52 | -32,400 | dd22feac7350a35f9d96c495e9a5cbb8748911a8 | wip (balanced ope) | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/__init__.py",
"new_path": "obp/ope/__init__.py",
"diff": "@@ -32,6 +32,7 @@ from obp.ope.meta import OffPolicyEvaluation\nfrom obp.ope.meta_continuous import ContinuousOffPolicyEvaluation\nfrom obp.ope.meta_slate import SlateOffPolicyEvaluation\nfrom o... | Python | Apache License 2.0 | st-tech/zr-obp | wip (balanced ope) |
641,006 | 29.11.2021 00:56:35 | -32,400 | 3257831052ad9aae82b2caa1b5ff8f7d1d4f3b18 | add estimated_pscore and fix past test | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/classification_model.py",
"new_path": "obp/ope/classification_model.py",
"diff": "@@ -261,7 +261,7 @@ class ImportanceSampler(BaseEstimator):\n`random_state` affects the ordering of the indices, which controls the randomness of each fold.\nSee https://... | Python | Apache License 2.0 | st-tech/zr-obp | add estimated_pscore and fix past test |
641,006 | 29.11.2021 14:06:15 | -32,400 | 0774cfd933f07edce9d313e5a5f2fa473855927a | fix test meta | [
{
"change_type": "MODIFY",
"old_path": "tests/ope/test_meta.py",
"new_path": "tests/ope/test_meta.py",
"diff": "@@ -402,6 +402,7 @@ def test_meta_create_estimator_inputs_using_valid_input_data(\n\"action_dist\",\n\"estimated_rewards_by_reg_model\",\n\"estimated_pscore\",\n+ \"importance_sampling_rat... | Python | Apache License 2.0 | st-tech/zr-obp | fix test meta |
641,014 | 19.12.2021 22:05:16 | 18,000 | dffcafbb020d081e4a71c13adbc54a95c8352e0f | implement SLOPE for hyperparam tuning of ope | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/estimators_tuning.py",
"new_path": "obp/ope/estimators_tuning.py",
"diff": "@@ -18,6 +18,7 @@ from .estimators import DoublyRobust\nfrom .estimators import DoublyRobustWithShrinkage\nfrom .estimators import InverseProbabilityWeighting\nfrom .estimators... | Python | Apache License 2.0 | st-tech/zr-obp | implement SLOPE for hyperparam tuning of ope |
641,006 | 31.12.2021 22:34:14 | -32,400 | eb9fa0de226e62aa473a01411ce5c0ec02f1a6bf | fix test of ipw and dr; add test of importance weight estimator; fix importance weight estimator | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/classification_model.py",
"new_path": "obp/ope/classification_model.py",
"diff": "@@ -13,7 +13,7 @@ from sklearn.utils import check_random_state\nfrom sklearn.utils import check_scalar\nfrom sklearn.calibration import CalibratedClassifierCV\n-from ..ut... | Python | Apache License 2.0 | st-tech/zr-obp | fix test of ipw and dr; add test of importance weight estimator; fix importance weight estimator |
641,006 | 01.01.2022 01:59:16 | -32,400 | 36d4d86d8f56b78e5565ea8d7fd68e116b9af582 | add tests of propensity score estimator | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/classification_model.py",
"new_path": "obp/ope/classification_model.py",
"diff": "@@ -287,7 +287,6 @@ class ImportanceWeightEstimator(BaseEstimator):\nif position is None or self.len_list == 1:\nposition = np.zeros_like(action)\nelse:\n- check_array(ar... | Python | Apache License 2.0 | st-tech/zr-obp | add tests of propensity score estimator |
641,006 | 01.01.2022 17:33:41 | -32,400 | dda71f05e5755c2758bb8df16ce46baca50159be | add bipw tests | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/estimators.py",
"new_path": "obp/ope/estimators.py",
"diff": "@@ -1810,13 +1810,13 @@ class DoublyRobustWithShrinkage(DoublyRobust):\nclass BalancedInverseProbabilityWeighting(BaseOffPolicyEstimator):\n\"\"\"Balanced Inverse Probability Weighting (IPW)... | Python | Apache License 2.0 | st-tech/zr-obp | add bipw tests |
641,006 | 01.01.2022 22:47:15 | -32,400 | 15080ddc9aafb59f77fdb5befed2111611999768 | add ope performance test | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/__init__.py",
"new_path": "obp/ope/__init__.py",
"diff": "@@ -82,6 +82,7 @@ __all_estimators__ = [\n\"DoublyRobustWithShrinkage\",\n\"SwitchDoublyRobust\",\n\"SelfNormalizedDoublyRobust\",\n+ \"BalancedInverseProbabilityWeighting\",\n]\n"
},
{
... | Python | Apache License 2.0 | st-tech/zr-obp | add ope performance test |
641,014 | 11.01.2022 06:53:32 | 18,000 | e7cac29b95f86d25976b322acadf59fd935aa1b2 | fix the weight of SGIPW/SGDR | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/estimators.py",
"new_path": "obp/ope/estimators.py",
"diff": "@@ -1830,7 +1830,7 @@ class SubGaussianInverseProbabilityWeighting(InverseProbabilityWeighting):\n\"\"\"\nn_rounds = action.shape[0]\niw = action_dist[np.arange(n_rounds), action, position] ... | Python | Apache License 2.0 | st-tech/zr-obp | fix the weight of SGIPW/SGDR |
641,006 | 11.01.2022 23:56:20 | -32,400 | ff27363251d9519dbb6a317f9c34f2d0152947f0 | Apply suggestions from code review
apply 2nd review | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/classification_model.py",
"new_path": "obp/ope/classification_model.py",
"diff": "@@ -18,8 +18,7 @@ from ..utils import check_array, sample_action_fast, check_bandit_feedback_input\n@dataclass\nclass ImportanceWeightEstimator(BaseEstimator):\n- \"\"\"M... | Python | Apache License 2.0 | st-tech/zr-obp | Apply suggestions from code review
apply 2nd review
Co-authored-by: yuta-saito <32621707+usaito@users.noreply.github.com> |
641,014 | 11.01.2022 12:46:04 | 18,000 | 11631614251eebf80c84e26d1722e44d087543e5 | implement synthetic data generator class for multiple loggers | [
{
"change_type": "MODIFY",
"old_path": "obp/dataset/__init__.py",
"new_path": "obp/dataset/__init__.py",
"diff": "@@ -11,6 +11,7 @@ from obp.dataset.synthetic import polynomial_behavior_policy\nfrom obp.dataset.synthetic import polynomial_reward_function\nfrom obp.dataset.synthetic import sparse_rew... | Python | Apache License 2.0 | st-tech/zr-obp | implement synthetic data generator class for multiple loggers |
641,006 | 12.01.2022 02:47:32 | -32,400 | 5ef06b30a43a78931713673d95626e84a290fbb3 | remove action_context testing when fitting method is raw | [
{
"change_type": "MODIFY",
"old_path": "tests/ope/test_importance_weight_estimator.py",
"new_path": "tests/ope/test_importance_weight_estimator.py",
"diff": "@@ -95,16 +95,6 @@ invalid_input_of_initializing_importance_weight_estimator = [\nValueError,\n\"fitting_method must be either 'sample' or 'ra... | Python | Apache License 2.0 | st-tech/zr-obp | remove action_context testing when fitting method is raw |
641,014 | 21.01.2022 22:15:19 | 18,000 | 18e4f7523cfc75704a7f172516c717bee342eb60 | add weighted dr | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/__init__.py",
"new_path": "obp/ope/__init__.py",
"diff": "@@ -27,6 +27,7 @@ from obp.ope.estimators_multi import MultiLoggersBalancedDoublyRobust\nfrom obp.ope.estimators_multi import MultiLoggersBalancedInverseProbabilityWeighting\nfrom obp.ope.estima... | Python | Apache License 2.0 | st-tech/zr-obp | add weighted dr |
641,014 | 21.01.2022 22:31:41 | 18,000 | f51b4698dcd14acdeed6f8b324a02f9b0827bce0 | rename synthetic multi loggers dataset class | [
{
"change_type": "MODIFY",
"old_path": "obp/dataset/__init__.py",
"new_path": "obp/dataset/__init__.py",
"diff": "@@ -18,7 +18,7 @@ from obp.dataset.synthetic_continuous import quadratic_reward_funcion_continuous\nfrom obp.dataset.synthetic_continuous import sign_synthetic_policy_continuous\nfrom ob... | Python | Apache License 2.0 | st-tech/zr-obp | rename synthetic multi loggers dataset class |
641,014 | 13.02.2022 12:13:51 | 18,000 | ef167582fbee65723faa00f1b7e443f25988847f | implement SyntheticBanditDatasetWithActionEmbeds | [
{
"change_type": "MODIFY",
"old_path": "obp/dataset/__init__.py",
"new_path": "obp/dataset/__init__.py",
"diff": "@@ -18,6 +18,7 @@ from obp.dataset.synthetic_continuous import quadratic_reward_funcion_continuous\nfrom obp.dataset.synthetic_continuous import sign_synthetic_policy_continuous\nfrom ob... | Python | Apache License 2.0 | st-tech/zr-obp | implement SyntheticBanditDatasetWithActionEmbeds |
641,014 | 13.02.2022 15:00:49 | 18,000 | d8c3d3846de8c762f2480e55a899c128ba4c6a43 | fix gradients | [
{
"change_type": "MODIFY",
"old_path": "obp/policy/offline.py",
"new_path": "obp/policy/offline.py",
"diff": "@@ -614,7 +614,7 @@ class QLearner(BaseOfflinePolicyLearner):\n@dataclass\nclass NNPolicyLearner(BaseOfflinePolicyLearner):\n- \"\"\"Off-policy learner parameterized by on a neural network.\... | Python | Apache License 2.0 | st-tech/zr-obp | fix gradients |
641,014 | 15.02.2022 07:53:40 | 18,000 | 5dd44315f043ece4878425ff353b7ed9e425524f | fix checkings | [
{
"change_type": "MODIFY",
"old_path": "obp/dataset/synthetic_embed.py",
"new_path": "obp/dataset/synthetic_embed.py",
"diff": "@@ -72,7 +72,7 @@ class SyntheticBanditDatasetWithActionEmbeds(SyntheticBanditDataset):\nn_cat_per_dim: int, default=10\nNumber of categories (cardinality) per category dim... | Python | Apache License 2.0 | st-tech/zr-obp | fix checkings |
641,014 | 17.02.2022 17:23:23 | 18,000 | 4932a4980b407e1378f0f9b20b495a422a214469 | add a parameter to control the embedding distribution for SyntheticBanditDatasetWithActionEmbeds | [
{
"change_type": "MODIFY",
"old_path": "obp/dataset/synthetic_embed.py",
"new_path": "obp/dataset/synthetic_embed.py",
"diff": "from dataclasses import dataclass\nfrom typing import Callable\nfrom typing import Optional\n+from typing import Union\nimport numpy as np\nfrom sklearn.utils import check_... | Python | Apache License 2.0 | st-tech/zr-obp | add a parameter to control the embedding distribution for SyntheticBanditDatasetWithActionEmbeds |
641,014 | 27.03.2022 07:02:46 | 14,400 | 66e052fb60359be849a1e07c8a83189bfb605fc7 | fix a bug of mipw with one-dimensional action embed | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/estimators_embed.py",
"new_path": "obp/ope/estimators_embed.py",
"diff": "@@ -293,6 +293,7 @@ class MarginalizedInverseProbabilityWeighting(BaseOffPolicyEstimator):\nelse:\ncheck_array(array=context, name=\"context\", expected_dim=2)\n+ if action_embed... | Python | Apache License 2.0 | st-tech/zr-obp | fix a bug of mipw with one-dimensional action embed |
641,014 | 27.03.2022 12:53:15 | 14,400 | a8a32782453db89ed9ebd96f85c4ce151dba04b7 | modify some variable names in MIPS | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/estimators_embed.py",
"new_path": "obp/ope/estimators_embed.py",
"diff": "@@ -44,7 +44,7 @@ class MarginalizedInverseProbabilityWeighting(BaseOffPolicyEstimator):\nn_actions: int\nNumber of actions in the logged data.\n- p_a_e_estimator: ClassifierMixi... | Python | Apache License 2.0 | st-tech/zr-obp | modify some variable names in MIPS |
641,014 | 02.04.2022 11:56:32 | 14,400 | 0dda2322adc8b38c75e02ead7e17a68521591a0f | modify init learning rate | [
{
"change_type": "MODIFY",
"old_path": "obp/policy/offline.py",
"new_path": "obp/policy/offline.py",
"diff": "@@ -743,7 +743,7 @@ class NNPolicyLearner(BaseOfflinePolicyLearner):\nsolver: str = \"adam\"\nalpha: float = 0.0001\nbatch_size: Union[int, str] = \"auto\"\n- learning_rate_init: float = 0.0... | Python | Apache License 2.0 | st-tech/zr-obp | modify init learning rate |
641,014 | 02.04.2022 11:56:57 | 14,400 | 22fd3adb0e6d41795a2893ad509bafc42f7259f7 | fix tests on mipw | [
{
"change_type": "MODIFY",
"old_path": "tests/ope/test_ipw_estimators_embed.py",
"new_path": "tests/ope/test_ipw_estimators_embed.py",
"diff": "@@ -314,9 +314,12 @@ def test_mipw_using_invalid_input_data(\n) -> None:\n# prepare ipw instances\nmipw = MIPW(n_actions=2)\n- snmipw = MIPW(n_actions=2)\n+... | Python | Apache License 2.0 | st-tech/zr-obp | fix tests on mipw |
641,014 | 03.04.2022 16:54:22 | 14,400 | 895e642bf1f18e37f9152b00f53e2a6ec7677dfb | refactor nnpolicylearner | [
{
"change_type": "MODIFY",
"old_path": "obp/policy/offline.py",
"new_path": "obp/policy/offline.py",
"diff": "@@ -1081,25 +1081,23 @@ class NNPolicyLearner(BaseOfflinePolicyLearner):\nself.nn_model.train()\nfor x, a, r, p, pos in training_data_loader:\noptimizer.zero_grad()\n- action_dist_by_current... | Python | Apache License 2.0 | st-tech/zr-obp | refactor nnpolicylearner |
641,014 | 03.04.2022 17:19:22 | 14,400 | 10b1f52dd02b1fa3ed2ef3222ed1e18783f404de | fix a bug in nnpolicylearner | [
{
"change_type": "MODIFY",
"old_path": "obp/policy/offline.py",
"new_path": "obp/policy/offline.py",
"diff": "@@ -1082,7 +1082,7 @@ class NNPolicyLearner(BaseOfflinePolicyLearner):\nfor x, a, r, p, pos in training_data_loader:\noptimizer.zero_grad()\npi = self.nn_model(x).unsqueeze(-1)\n- policy_gra... | Python | Apache License 2.0 | st-tech/zr-obp | fix a bug in nnpolicylearner |
641,014 | 03.04.2022 18:22:41 | 14,400 | 52010ce2ff38ea01082bf69f1387c7dd9a6e5740 | modify def init_learning_rate of nnpolicylearner | [
{
"change_type": "MODIFY",
"old_path": "obp/policy/offline.py",
"new_path": "obp/policy/offline.py",
"diff": "@@ -743,7 +743,7 @@ class NNPolicyLearner(BaseOfflinePolicyLearner):\nsolver: str = \"adam\"\nalpha: float = 0.0001\nbatch_size: Union[int, str] = \"auto\"\n- learning_rate_init: float = 0.0... | Python | Apache License 2.0 | st-tech/zr-obp | modify def init_learning_rate of nnpolicylearner |
641,014 | 03.04.2022 18:22:58 | 14,400 | 21b5da2f5fccf9380f5802b1c35bc877f48c1f02 | update example notebooks | [
{
"change_type": "MODIFY",
"old_path": "examples/quickstart/multiclass.ipynb",
"new_path": "examples/quickstart/multiclass.ipynb",
"diff": "},\n{\n\"cell_type\": \"code\",\n- \"execution_count\": 2,\n+ \"execution_count\": 1,\n\"metadata\": {},\n\"outputs\": [],\n\"source\": [\n},\n{\n\"cell_type\":... | Python | Apache License 2.0 | st-tech/zr-obp | update example notebooks |
641,014 | 04.04.2022 14:56:53 | 14,400 | 820a5792eb3b842991435c76f204f376a22d90e2 | run with a new version | [
{
"change_type": "MODIFY",
"old_path": "examples/quickstart/multiclass.ipynb",
"new_path": "examples/quickstart/multiclass.ipynb",
"diff": "},\n{\n\"cell_type\": \"code\",\n- \"execution_count\": 1,\n+ \"execution_count\": 2,\n\"metadata\": {},\n\"outputs\": [],\n\"source\": [\n},\n{\n\"cell_type\":... | Python | Apache License 2.0 | st-tech/zr-obp | run with a new version |
641,014 | 24.04.2022 13:10:22 | 14,400 | 87497eba440d3ae7e0a14a487309acf61ee50e85 | automatic hyperparam sort for slope | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/estimators.py",
"new_path": "obp/ope/estimators.py",
"diff": "@@ -526,11 +526,11 @@ class InverseProbabilityWeighting(BaseOffPolicyEstimator):\n(If only a single action is chosen for each data, you can just ignore this argument.)\nuse_bias_upper_bound:... | Python | Apache License 2.0 | st-tech/zr-obp | automatic hyperparam sort for slope |
641,014 | 24.04.2022 14:33:56 | 14,400 | 9ee8704af78dc73fdbf16d79624d1ef33a57a578 | add some off policy gradient estimators | [
{
"change_type": "MODIFY",
"old_path": "obp/policy/offline.py",
"new_path": "obp/policy/offline.py",
"diff": "@@ -348,7 +348,7 @@ class IPWLearner(BaseOfflinePolicyLearner):\n@dataclass\nclass QLearner(BaseOfflinePolicyLearner):\n- \"\"\"Off-policy learner based on Direct Method.\n+ \"\"\"Off-policy... | Python | Apache License 2.0 | st-tech/zr-obp | add some off policy gradient estimators |
641,014 | 24.04.2022 21:26:48 | 14,400 | 437ab5cce135f2fb0b7f6608c6909cbe4eef8416 | fix the automatic hyperparam sorting of dr-os | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/estimators_tuning.py",
"new_path": "obp/ope/estimators_tuning.py",
"diff": "@@ -44,7 +44,7 @@ class BaseOffPolicyEstimatorTuning:\nWhether to use a bias upper bound in hyperparameter tuning.\nIf False, the direct bias estimator is used to estimate the ... | Python | Apache License 2.0 | st-tech/zr-obp | fix the automatic hyperparam sorting of dr-os |
641,014 | 24.04.2022 22:35:23 | 14,400 | 02493699e0cdf27963a76d8110d1df6c6b0389f5 | fix added policy gradients | [
{
"change_type": "MODIFY",
"old_path": "obp/policy/offline.py",
"new_path": "obp/policy/offline.py",
"diff": "@@ -796,6 +796,8 @@ class NNPolicyLearner(BaseOfflinePolicyLearner):\n)\nif self.off_policy_objective == \"ipw-subgauss\":\n+ if self.lambda_ is None:\n+ self.lambda_ = 0.001\ncheck_scalar(\... | Python | Apache License 2.0 | st-tech/zr-obp | fix added policy gradients |
641,014 | 28.04.2022 16:59:32 | 14,400 | 7f2ab7857ad25f9f9a20d302369bd0b3e95af397 | fix error checkings on p_e_a | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/meta.py",
"new_path": "obp/ope/meta.py",
"diff": "@@ -149,15 +149,26 @@ class OffPolicyEvaluation:\npass\nelif isinstance(value_or_dict, dict):\nfor estimator_name, value in value_or_dict.items():\n+ expected_dim = 1\n+ if var_name in [\"p_e_a\", \"pi_... | Python | Apache License 2.0 | st-tech/zr-obp | fix error checkings on p_e_a |
641,011 | 10.05.2022 10:15:46 | -32,400 | cb80f49fd448dfe85e29dcfead848c93a404d879 | fix expected reward factual in indepedent case | [
{
"change_type": "MODIFY",
"old_path": "obp/dataset/synthetic_slate.py",
"new_path": "obp/dataset/synthetic_slate.py",
"diff": "@@ -1441,7 +1441,7 @@ def action_interaction_reward_function(\naction_2d[:, pos_],\n]\nif reward_structure == \"independent\":\n- continue\n+ pass\nelif is_additive:\nfor p... | Python | Apache License 2.0 | st-tech/zr-obp | fix expected reward factual in indepedent case |
641,014 | 15.06.2022 12:43:03 | -32,400 | 83f3945f3060b59dd2ca1f2995f8eadb0bacb992 | allow slope to use the true maginal importance weight for mips | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/estimators_embed.py",
"new_path": "obp/ope/estimators_embed.py",
"diff": "@@ -303,6 +303,7 @@ class MarginalizedInverseProbabilityWeighting(BaseOffPolicyEstimator):\nposition=position,\npi_b=pi_b,\naction_dist=action_dist,\n+ p_e_a=p_e_a,\n)\nelif self... | Python | Apache License 2.0 | st-tech/zr-obp | allow slope to use the true maginal importance weight for mips |
641,014 | 15.06.2022 13:22:16 | -32,400 | 0e94113752859305765e1ceebbc096ca6da8873d | fix a bug in slope in mips | [
{
"change_type": "MODIFY",
"old_path": "obp/ope/estimators_embed.py",
"new_path": "obp/ope/estimators_embed.py",
"diff": "@@ -347,6 +347,18 @@ class MarginalizedInverseProbabilityWeighting(BaseOffPolicyEstimator):\ncomb_list = list(itertools.combinations(feat_list, i))\ntheta_list_, cnf_list_ = [], ... | Python | Apache License 2.0 | st-tech/zr-obp | fix a bug in slope in mips |
641,008 | 29.06.2022 16:36:51 | -7,200 | f7918ad8d4641ee2dd1a99050c9751cb68852641 | Add factual rewards to synthetic data generation and enable online bandit simulations | [
{
"change_type": "MODIFY",
"old_path": "obp/dataset/synthetic.py",
"new_path": "obp/dataset/synthetic.py",
"diff": "@@ -237,14 +237,16 @@ class SyntheticBanditDataset(BaseBanditDataset):\ndef sample_reward_given_expected_reward(\nself,\nexpected_reward: np.ndarray,\n- action: np.ndarray,\n+ action: ... | Python | Apache License 2.0 | st-tech/zr-obp | Add factual rewards to synthetic data generation and enable online bandit simulations |
641,008 | 11.07.2022 16:24:25 | -7,200 | 9f39574e06488982eceaf00118ac25a9cddff690 | reward independent delays for the simulator | [
{
"change_type": "MODIFY",
"old_path": "obp/simulator/simulator.py",
"new_path": "obp/simulator/simulator.py",
"diff": "\"\"\"Bandit Simulator.\"\"\"\nfrom copy import deepcopy\n-from typing import Callable\n+from typing import Callable, Dict, List\nfrom typing import Union\nimport numpy as np\n+imp... | Python | Apache License 2.0 | st-tech/zr-obp | reward independent delays for the simulator |
641,008 | 12.07.2022 11:37:51 | -7,200 | b5f9bd42cd5523b48e24ae553ba6fbe59afa9774 | reward depend delays for the simulators | [
{
"change_type": "MODIFY",
"old_path": "obp/dataset/synthetic.py",
"new_path": "obp/dataset/synthetic.py",
"diff": "@@ -358,7 +358,11 @@ class SyntheticBanditDataset(BaseBanditDataset):\nround_delays = None\nif self.delay_function:\n- round_delays = self.delay_function(actions.shape[0])\n+ round_del... | Python | Apache License 2.0 | st-tech/zr-obp | reward depend delays for the simulators |
641,008 | 12.07.2022 16:02:02 | -7,200 | 9c53a193fc349828c0f97cf5830ca150571af841 | psuedocode and coefficient adjustments for drift implemetnation | [
{
"change_type": "MODIFY",
"old_path": "obp/dataset/synthetic.py",
"new_path": "obp/dataset/synthetic.py",
"diff": "\"\"\"Class for Generating Synthetic Logged Bandit Data.\"\"\"\nfrom dataclasses import dataclass\n-from typing import Callable\n+from typing import Callable, Tuple\nfrom typing import... | Python | Apache License 2.0 | st-tech/zr-obp | psuedocode and coefficient adjustments for drift implemetnation |
641,008 | 14.07.2022 18:03:26 | -7,200 | 3f2eb265ab08198e3412002521722f282c7a6758 | Added transition period, transition style, seasonality and base coef to drift | [
{
"change_type": "MODIFY",
"old_path": "obp/dataset/synthetic.py",
"new_path": "obp/dataset/synthetic.py",
"diff": "# Licensed under the Apache 2.0 License.\n\"\"\"Class for Generating Synthetic Logged Bandit Data.\"\"\"\n+from collections import deque\nfrom dataclasses import dataclass\nfrom typing... | Python | Apache License 2.0 | st-tech/zr-obp | Added transition period, transition style, seasonality and base coef to drift |
641,008 | 21.07.2022 14:02:24 | -7,200 | a189fe29faf00847ccb0bfd5e7b7c2c71a0f2faf | working state for drift, tests are failing | [
{
"change_type": "MODIFY",
"old_path": "obp/dataset/synthetic.py",
"new_path": "obp/dataset/synthetic.py",
"diff": "@@ -70,13 +70,15 @@ class CoefficientDrifter:\nbase_coefficient_weight: float = 0.0\neffective_dim_action_context: Optional[int] = None\neffective_dim_context: Optional[int] = None\n- ... | Python | Apache License 2.0 | st-tech/zr-obp | working state for drift, tests are failing |
641,008 | 26.07.2022 15:41:35 | -7,200 | 600d7af8890356a9729b85dc97584b884395ce93 | speed up simulation by not doing constant vstacks | [
{
"change_type": "MODIFY",
"old_path": "obp/simulator/simulator.py",
"new_path": "obp/simulator/simulator.py",
"diff": "@@ -377,12 +377,23 @@ class BanditPolicySimulator:\nfor _ in tqdm(range(n_rounds)):\nself.step()\nif batch_bandit_rounds:\n+ start_round = self.rounds_played\n+ try:\n+ # Append co... | Python | Apache License 2.0 | st-tech/zr-obp | speed up simulation by not doing constant vstacks |
641,008 | 26.07.2022 17:01:56 | -7,200 | 74e9a9f92c397d866e48af995923ba179bdec91b | remove accidental code duplication and style changes | [
{
"change_type": "MODIFY",
"old_path": "obp/simulator/simulator.py",
"new_path": "obp/simulator/simulator.py",
"diff": "@@ -390,12 +390,14 @@ class BanditPolicySimulator:\n# If anything goes wrong, we want to remove all contexts and rewards that have not yet been shown yet\ntotal_rounds = batch_band... | Python | Apache License 2.0 | st-tech/zr-obp | remove accidental code duplication and style changes |
641,008 | 01.08.2022 11:25:15 | -7,200 | ff54d195f8c155bae537807bb2bbad742fc7aa38 | speed up simulation by using python lists instead of constant numpy appends | [
{
"change_type": "MODIFY",
"old_path": "obp/simulator/simulator.py",
"new_path": "obp/simulator/simulator.py",
"diff": "@@ -50,7 +50,7 @@ class BanditRounds:\nround_delays: np.ndarray\ndef _get_bandit_round(self) -> BanditRound:\n- if np.any(self.round_delays):\n+ if self.round_delays is not None:\n... | Python | Apache License 2.0 | st-tech/zr-obp | speed up simulation by using python lists instead of constant numpy appends |
641,008 | 11.08.2022 13:43:45 | -7,200 | 969b60630a96ee3213b03d5c382102f3b98ba93b | optional z-score | [
{
"change_type": "MODIFY",
"old_path": "obp/dataset/synthetic.py",
"new_path": "obp/dataset/synthetic.py",
"diff": "@@ -113,7 +113,7 @@ class SyntheticBanditDataset(BaseBanditDataset):\n.. code-block:: python\n- >>> from obp.dataset import (\n+ >>> from obp.env import (\nSyntheticBanditDataset,\nlog... | Python | Apache License 2.0 | st-tech/zr-obp | optional z-score |
641,008 | 18.09.2022 15:29:24 | 25,200 | ad015d18b99214e29f950ab140233b77e36af04e | addressed comments related to naming and addition of argument docstring for the coefficient drifter | [
{
"change_type": "RENAME",
"old_path": "examples/quickstart/online-vs-ope-simulation.ipynb",
"new_path": "examples/quickstart/online-bandit-vs-opl-simulation.ipynb",
"diff": ""
},
{
"change_type": "RENAME",
"old_path": "examples/quickstart/delay-simulation.ipynb",
"new_path": "exampl... | Python | Apache License 2.0 | st-tech/zr-obp | addressed comments related to naming and addition of argument docstring for the coefficient drifter |
641,008 | 22.09.2022 11:42:08 | 25,200 | d4be5c5bc18537a5b4fb349a95598f52d74e6b27 | Renamed online to replay through the examples and README | [
{
"change_type": "MODIFY",
"old_path": "examples/README.md",
"new_path": "examples/README.md",
"diff": "@@ -5,6 +5,6 @@ This page contains a list of examples written with Open Bandit Pipeline.\n- [`obd/`](./obd/): example implementations for evaluating standard off-policy estimators with the small s... | Python | Apache License 2.0 | st-tech/zr-obp | Renamed online to replay through the examples and README |
641,008 | 03.11.2022 16:40:17 | -3,600 | 97980a3087d00647bf104e8faac6bb342344b5c8 | more robust tests by removing fixed values and increasing rounds | [
{
"change_type": "MODIFY",
"old_path": "tests/dataset/test_synthetic.py",
"new_path": "tests/dataset/test_synthetic.py",
"diff": "@@ -581,12 +581,5 @@ def test_base_reward_create_a_matrix_with_expected_rewards_with_identical_expect\ncontext, action_context, degree=5, effective_dim_ratio=1.0, random_... | Python | Apache License 2.0 | st-tech/zr-obp | more robust tests by removing fixed values and increasing rounds |
641,008 | 04.11.2022 13:31:27 | -3,600 | abd9206822a571ace0d68198aa1d5054eca89d3c | run black for linting | [
{
"change_type": "MODIFY",
"old_path": "tests/ope/conftest.py",
"new_path": "tests/ope/conftest.py",
"diff": "@@ -157,6 +157,7 @@ def feedback_key_set() -> Set[str]:\n\"reward\",\n}\n+\n# random evaluation policy\n@pytest.fixture(scope=\"session\")\ndef random_action_dist(synthetic_bandit_feedback) ... | Python | Apache License 2.0 | st-tech/zr-obp | run black for linting |
49,717 | 03.01.2017 16:44:28 | 28,800 | 6963f5e10a7c3e462b24013f1a35cc656718f2dd | Hotfix to make sure only uak+/ua+ is compiled for GPU.
Closes | [
{
"change_type": "MODIFY",
"old_path": "src/main/java/org/apache/sysml/hops/AggUnaryOp.java",
"new_path": "src/main/java/org/apache/sysml/hops/AggUnaryOp.java",
"diff": "@@ -145,9 +145,12 @@ public class AggUnaryOp extends Hop implements MultiThreadedHop\nelse { //general case\nint k = OptimizerUtil... | Java | Apache License 2.0 | apache/systemds | Hotfix to make sure only uak+/ua+ is compiled for GPU.
Closes #325. |
49,736 | 06.01.2017 10:20:30 | 28,800 | 4cd98291780691082be07b1a817bd71c1024ed62 | Support bias_add operation
Also added an external UDF for faster SGD Nesterov's update. However
since the performance improvement was only by 7%, I decided not to
introduce a fused operator for it. We can revisit this in a later PR.
bias_add should work for both CP and GPU.
Closes | [
{
"change_type": "MODIFY",
"old_path": "scripts/staging/SystemML-NN/nn/layers/conv_builtin.dml",
"new_path": "scripts/staging/SystemML-NN/nn/layers/conv_builtin.dml",
"diff": "@@ -60,6 +60,7 @@ forward = function(matrix[double] X, matrix[double] W, matrix[double] b,\n*/\nN = nrow(X)\nF = nrow(W)\n+ ... | Java | Apache License 2.0 | apache/systemds | [SYSTEMML-540] Support bias_add operation
- Also added an external UDF for faster SGD Nesterov's update. However
since the performance improvement was only by 7%, I decided not to
introduce a fused operator for it. We can revisit this in a later PR.
- bias_add should work for both CP and GPU.
Closes #328. |
49,736 | 06.01.2017 10:23:47 | 28,800 | 2b5b12557556d95c499730ae105807d996ad7a47 | Added fused relu_maxpooling
Fused relu_maxpooling reduces the unnecessary dense-to-sparse-to-dense
conversion. This operator makes relu a "no op".
Note: fused relu_maxpooling is only supported in CP, not on GPU as both
relu and maxpooling invoke CuDNN functions.
Also, improved the performance of maxpooling by computing indexes
apriori.
Closes | [
{
"change_type": "MODIFY",
"old_path": "src/main/java/org/apache/sysml/hops/ConvolutionOp.java",
"new_path": "src/main/java/org/apache/sysml/hops/ConvolutionOp.java",
"diff": "@@ -25,6 +25,7 @@ import org.apache.sysml.api.DMLScript;\nimport org.apache.sysml.conf.ConfigurationManager;\nimport org.apa... | Java | Apache License 2.0 | apache/systemds | [SYSTEMML-540] Added fused relu_maxpooling
- Fused relu_maxpooling reduces the unnecessary dense-to-sparse-to-dense
conversion. This operator makes relu a "no op".
- Note: fused relu_maxpooling is only supported in CP, not on GPU as both
relu and maxpooling invoke CuDNN functions.
- Also, improved the performance of maxpooling by computing indexes
apriori.
Closes #329. |
49,736 | 06.01.2017 10:28:54 | 28,800 | 6f8cea9bc2d42913d0cf1c917c1aada9aa55bee1 | Add support for cublas daxpy operation
Closes | [
{
"change_type": "MODIFY",
"old_path": "src/main/java/org/apache/sysml/hops/TernaryOp.java",
"new_path": "src/main/java/org/apache/sysml/hops/TernaryOp.java",
"diff": "package org.apache.sysml.hops;\n+import org.apache.sysml.api.DMLScript;\nimport org.apache.sysml.conf.ConfigurationManager;\nimport ... | Java | Apache License 2.0 | apache/systemds | [SYSTEMML-446] Add support for cublas daxpy operation
Closes #330. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.