1171f9e | Lorenzo Stella | 21 January 2020, 10:02:22 UTC | fix item_id field in provided datasets (#566) | 21 January 2020, 10:02:22 UTC |
824159c | Kashif Rasul | 20 January 2020, 21:28:30 UTC | fixed taxi dataset cardinality (#552) * fixed taxi dataset cardinality * Update _gp_copula_2019.py Points now to the corrected taxi file. * Update _gp_copula_2019.py Fix style check * Update _gp_copula_2019.py Minor. * Update _gp_copula_2019.py Style fix Co-authored-by: Michael Bohlke-Schneider <mschneid83@gmail.com> | 20 January 2020, 21:28:29 UTC |
6338eb2 | Lorenzo Stella | 20 January 2020, 15:10:13 UTC | Fix bug in cdf method of piecewise linear distributions (#564) * fix bug in cdf method of piecewise linear distributions * improve test * regularize slopes and add robustness test | 20 January 2020, 15:10:13 UTC |
55ee968 | Jasper Schulz | 16 January 2020, 18:58:16 UTC | Removed timeouts from tests. Added hard timeout of 30seconds for all tests. (#562) | 16 January 2020, 18:58:16 UTC |
c58cf39 | Lorenzo Stella | 15 January 2020, 17:02:14 UTC | remove dataset class in favor of alias (#560) | 15 January 2020, 17:02:14 UTC |
d647aa4 | Lorenzo Stella | 10 January 2020, 07:00:02 UTC | prob -> log_prob (#556) | 10 January 2020, 07:00:02 UTC |
753add5 | Micky C | 07 January 2020, 23:17:53 UTC | add poisson distribution (#532) Co-authored-by: Lorenzo Stella <lorenzostella@gmail.com> | 07 January 2020, 23:17:53 UTC |
d5a52c7 | Aaron Spieler | 07 January 2020, 13:49:40 UTC | Distribution dependent loss masking. (#534) * Loss is now masked with F.where for unobserved values. * Value in support exposed through DistributionOutput. Co-authored-by: Lorenzo Stella <lorenzostella@gmail.com> | 07 January 2020, 13:49:40 UTC |
0bfa906 | Kashif Rasul | 07 January 2020, 12:38:47 UTC | Shopping holidays (#542) * shopping holidays added black friday and cyber monday * fix for #469 fixes #469 by adding 1/2 a week more days * formatting * use Day | 07 January 2020, 12:38:47 UTC |
a4e5584 | Lorenzo Stella | 06 January 2020, 16:53:32 UTC | Add `loc` argument to distribution output classes (#540) * add loc to distribution output classes * fix docstring | 06 January 2020, 16:53:32 UTC |
4bf7f52 | Valentin Flunkert | 02 January 2020, 10:25:15 UTC | fix quantile fn for transformed distribution (#544) | 02 January 2020, 10:25:15 UTC |
5c15eb6 | Jasper Schulz | 20 December 2019, 14:39:18 UTC | Quickfix for SageMaker SDK. | 20 December 2019, 14:58:29 UTC |
0e1bf3d | Dr. Kashif Rasul | 20 December 2019, 14:41:04 UTC | Fixed FileDataset SourceContext (#538) Source context's source is the path rather than the whole tuple. | 20 December 2019, 14:41:04 UTC |
a6d09a1 | Syama Sundar Rangapuram | 19 December 2019, 10:26:54 UTC | Fix #535: quantile of Binned distribution (#536) | 19 December 2019, 10:26:54 UTC |
1e930e5 | Aaron Spieler | 18 December 2019, 10:40:07 UTC | Fixed distribution sampling issues. (#526) * Fixed distribution invalid sampling issues. Added machine epsilon as minimal delta. * Using default dtype now. | 18 December 2019, 10:40:07 UTC |
a9cde7b | Lorenzo Stella | 17 December 2019, 14:55:19 UTC | Fix instance splitter issue with short time series (#533) * fix instance splitter issue with short time series * improve code * add test cases * address comments | 17 December 2019, 14:55:19 UTC |
0213131 | Aaron Spieler | 12 December 2019, 13:55:04 UTC | Sagemaker SDK Integration (#444) * Added Docker file for CPU with gluonts preinstalled. | 12 December 2019, 13:55:04 UTC |
be71253 | Aaron Spieler | 11 December 2019, 21:10:58 UTC | Beta distribution implemented. (#512) * Beta distribution implemented. * Dirichlet comment removed Co-Authored-By: Lorenzo Stella <lorenzostella@gmail.com> * Test changed to .5 instead of close to 1 Co-Authored-By: Lorenzo Stella <lorenzostella@gmail.com> * Blacked the file. | 11 December 2019, 21:10:58 UTC |
08ab434 | Aaron Spieler | 10 December 2019, 13:37:34 UTC | Gamma distribution implemented. (#502) * Gamma distribution implemented. * Remove forgotten comment. Co-Authored-By: Lorenzo Stella <lorenzostella@gmail.com> * Fixed issue 287 with respect to the Gamma distribution. | 10 December 2019, 13:37:34 UTC |
a69e397 | Jasper Schulz | 10 December 2019, 09:09:00 UTC | Removed unused function rollback. (#506) | 10 December 2019, 09:09:00 UTC |
c9bd5fd | Jasper Schulz | 09 December 2019, 16:24:14 UTC | Added operator to allow for easier chaining of transformations. (#505) | 09 December 2019, 16:24:14 UTC |
fe065b4 | Jasper Schulz | 09 December 2019, 14:40:26 UTC | Split up transform into its own module. (#499) * Added chain method to Transformation. | 09 December 2019, 14:40:26 UTC |
4bc1f36 | Jasper Schulz | 09 December 2019, 12:49:18 UTC | Added strategies for timezone information. (#500) * Added strategies for timezone information. * Applied black. * Quickfix. | 09 December 2019, 12:49:18 UTC |
5f34790 | Dr. Kashif Rasul | 09 December 2019, 10:44:52 UTC | Fixed forecast test assert string. (#504) | 09 December 2019, 10:44:52 UTC |
fbcb244 | Caner Turkmen | 07 December 2019, 09:44:25 UTC | add transform objects for temporal point processes (#294) (#341) | 07 December 2019, 09:44:25 UTC |
9199611 | Jasper Schulz | 06 December 2019, 15:32:09 UTC | Changed FileDataset to be more easily inheritable. (#498) | 06 December 2019, 15:32:09 UTC |
0bdfda6 | Michael Bohlke-Schneider | 06 December 2019, 12:02:40 UTC | Add GP-Copula model. (#497) * Starting with GPVAR port * Starting with GPVAR port * Changes for GP model * Added GP Copula model with smoke tests * Fixes for GPVAR model * Fixed typo * Changed initialization in lowrank multivariate gaussian inference test * Increase time for lowrank multivariate test to fix build on windows machines * Included missing headers * Removed unused import * Increased ND for GPVAR smoke tests | 06 December 2019, 12:02:40 UTC |
5ffce32 | Michael Bohlke-Schneider | 04 December 2019, 14:30:11 UTC | Introduces DeepVAR model (#491) * Added helper functions used by DeepVAR and GPVar models * Added DeepVAR * Added test for DeepVAR * Changed DeepVAR model to NTC output * Some refactoring on the DeepVAR network * Addressed black and mypy complains * Added standard header to test_deepvar * Renaming and refactoring to make class descriptions consistent * More refactoring * Added missing headers * Removed replaced get_granularity with buildin pandas function * Refactored distribution output construction in DeepVAR * Test refactored default distribution output for DeepVAR * Renamed argument names in DeepVAR test for consistency | 04 December 2019, 14:30:11 UTC |
913c91e | Dr. Kashif Rasul | 03 December 2019, 16:22:08 UTC | fix typo in test (#492) | 03 December 2019, 16:22:08 UTC |
4eab4df | Michael Bohlke-Schneider | 02 December 2019, 10:22:35 UTC | Marginal CDFtoGaussianTransformation (#486) * Ported over marginal transformation code * Added monkey patched take_along_axis operator that we can use until we are able to bump up numpy version * Templated erf/erfinf operator such that it works with nd/mx.sym/mx.nd * Refactored marginal CDFtoGaussianTransformation * Added test for gaussian CDF/PDF in CDFToGaussianTransformation * Addressed PR comments | 02 December 2019, 10:22:35 UTC |
385afd8 | Michael Bohlke-Schneider | 29 November 2019, 18:50:36 UTC | Added GP Copula datasets (#476) | 29 November 2019, 18:50:36 UTC |
829e3a1 | FadhelA | 28 November 2019, 15:45:12 UTC | Dirichlet Multinomial distribution (#482) * Creating dirichlet-multinomial distribution * loss sample and properties of dirichlet multinomial * tests for dirichlet multinomial * tests for dirichlet multinomial * Creating dirichlet-multinomial distribution * loss sample and properties of dirichlet multinomial * tests for dirichlet multinomial * tests for dirichlet multinomial * test for dirichlet multinomial * test for dirichlet multinomial * fix test inference | 28 November 2019, 15:45:12 UTC |
58e2937 | Leonard Lausen | 28 November 2019, 13:46:55 UTC | Specify --upgrade in README.md install example (#480) Try to avoid user problems such as https://github.com/apache/incubator-mxnet/issues/16924 | 28 November 2019, 13:46:55 UTC |
596123a | Lorenzo Stella | 28 November 2019, 10:04:25 UTC | tranform -> transform (#484) | 28 November 2019, 10:04:25 UTC |
b9ee9cb | Lorenzo Stella | 27 November 2019, 10:14:30 UTC | forcecast -> forecast (#479) | 27 November 2019, 10:14:30 UTC |
d257f7f | Jasper Schulz | 26 November 2019, 21:22:42 UTC | Update README.md (#478) Make PyPi version link to https://pypi.org/project/gluonts/ | 26 November 2019, 21:22:42 UTC |
d37f057 | Jasper Schulz | 26 November 2019, 16:22:43 UTC | Refined doc requiremnents; using sphinx 2. (#477) | 26 November 2019, 16:22:43 UTC |
6fd367e | Syama Sundar Rangapuram | 25 November 2019, 20:51:37 UTC | fix validation error in transformed distribution (#475) | 25 November 2019, 20:51:37 UTC |
ca8d14d | Jasper Schulz | 21 November 2019, 13:53:39 UTC | Remove warning in tests. (#473) | 21 November 2019, 13:53:39 UTC |
c275e98 | Jasper Schulz | 21 November 2019, 13:42:28 UTC | Fix use of numpy.histogram. (#472) `normed` is deprecated in favour of `density`. | 21 November 2019, 13:42:28 UTC |
3e38610 | Jasper Schulz | 21 November 2019, 13:35:06 UTC | Added mxnet nightly builds. (#471) | 21 November 2019, 13:35:06 UTC |
2cb59f4 | Michael Bohlke-Schneider | 20 November 2019, 17:11:17 UTC | Added aggregation functionality to MultivariateEvaluator (#459) * Added aggregation evaluations in MultivariateEvaluator * Made black happy * Addressed PR comments * Even more type fixes * Trying to make the doc build happy | 20 November 2019, 17:11:17 UTC |
730705e | Stephan Rabanser | 20 November 2019, 16:43:17 UTC | Fix WaveNet prediction length during training (#347) * Fix WaveNet prediction length during training * Fix log text * Add prediction length guard * Default training prediction length to inference prediction length * Set training window length default value based on prediction length * Fix InstanceSplitter prediction length based on training window * Fix WaveNet training prediction for good | 20 November 2019, 16:43:17 UTC |
7851a50 | Jasper Schulz | 20 November 2019, 10:58:06 UTC | Update docs.yml (#466) Added doctests. | 20 November 2019, 10:58:06 UTC |
87d7ada | Jasper Schulz | 20 November 2019, 10:48:42 UTC | Updated pydantic to version 1. (#465) * Updated pydantic to version 1. This required some renaming of some fields. * Also updated requirements.txt with `~=` as version specifier. | 20 November 2019, 10:48:42 UTC |
9da9dcd | Tom Rochette | 20 November 2019, 10:19:34 UTC | Relax requirements constraints (#456) This allows gluon-ts to be installed in a project that requires similar dependencies. Because gluon-ts dependencies constraints are too strict, it may prevent it from being installed side by side with those other libraries if the constraints cannot resolve to a common version. | 20 November 2019, 10:19:34 UTC |
b1bed75 | Dr. Kashif Rasul | 19 November 2019, 09:27:26 UTC | Removed unused static method in DeepARNetwork (#460) | 19 November 2019, 09:27:26 UTC |
db99013 | Jasper Schulz | 14 November 2019, 15:29:59 UTC | Added median as alias for p50 to Forecast. (#450) | 14 November 2019, 15:29:59 UTC |
450ae3f | Anton Kiselev | 14 November 2019, 09:17:25 UTC | Use validation to prevent overfitting (#378) * Use validation set to prevent overfitting * Review changes for #378 * Using InferenceDataLoader for validation * Typo in docs * Using ValidationDataLoader instead of InferenceDataLoader. Using "validation" instead of "val" and "valid" | 14 November 2019, 09:17:25 UTC |
a9f00a1 | Lorenzo Stella | 13 November 2019, 17:26:11 UTC | move warning to the right place (#447) | 13 November 2019, 17:26:11 UTC |
11fed53 | Lorenzo Stella | 13 November 2019, 12:57:37 UTC | fix deepstate serialization, add tests (#445) | 13 November 2019, 12:57:37 UTC |
85ea218 | Jasper Schulz | 11 November 2019, 18:16:15 UTC | Serde cleanup. (#441) * Serde cleanup. | 11 November 2019, 18:16:15 UTC |
34167de | Jasper Schulz | 11 November 2019, 17:10:51 UTC | Added docs to Actions. (#442) * Added docs to Actions. | 11 November 2019, 17:10:51 UTC |
320d2b7 | Jasper Schulz | 11 November 2019, 15:44:01 UTC | Added flaky to test_multivariate_sampling. (#440) In some cases np.allclose fails. | 11 November 2019, 15:44:01 UTC |
312f45b | Jasper Schulz | 11 November 2019, 14:37:18 UTC | Added windows to CI python build. (#438) * Added windows to CI python build. * Skip shell tests on windows. * Fixed botocore requirement. | 11 November 2019, 14:37:18 UTC |
36fe4c0 | Jasper Schulz | 11 November 2019, 14:25:25 UTC | Fix escaping of string in serde.dump_code. (#439) | 11 November 2019, 14:25:25 UTC |
9f65489 | Michael Bohlke-Schneider | 11 November 2019, 10:22:26 UTC | Added multivariate grouper and tests (#432) * Added multivariate grouper and tests * Refactoring of multivariate grouper * Addressing PR comments. * Fixed documention | 11 November 2019, 10:22:26 UTC |
f5615d5 | Lorenzo Stella | 11 November 2019, 10:07:40 UTC | Fixes to setup.py to make it work on Windows (#433) | 11 November 2019, 10:07:40 UTC |
7cc3b54 | Lorenzo Stella | 10 November 2019, 07:22:57 UTC | remove some occurrences of num_eval_samples (#436) | 10 November 2019, 07:22:57 UTC |
049d08a | Timo Meiendresch | 09 November 2019, 13:27:28 UTC | Update benchmark_m4.py (#435) Incorporate change in #421 to this example. | 09 November 2019, 13:27:28 UTC |
50b3ef3 | Lorenzo Stella | 08 November 2019, 18:41:16 UTC | Update README.md (#434) | 08 November 2019, 18:41:16 UTC |
4da2167 | Syama Sundar Rangapuram | 08 November 2019, 14:34:24 UTC | DeepStateEstimator: Some adjustments to hyperparameter settings (#415) * DeepStateEstimator: Some adjustments to hyperparameter settings * fix issues with import * remove empty line * address comments * fixing bugs * address comment | 08 November 2019, 14:34:24 UTC |
a705cce | Lorenzo Stella | 08 November 2019, 10:20:58 UTC | remove mxnet requirement (#429) | 08 November 2019, 10:20:58 UTC |
af42ef0 | Konstantinos Benidis | 08 November 2019, 09:56:53 UTC | num_sample variation fixes (#421) * fix num_sample variations | 08 November 2019, 09:56:53 UTC |
d8f56f7 | Lorenzo Stella | 07 November 2019, 14:35:24 UTC | Update dirichlet.py (#425) | 07 November 2019, 14:35:24 UTC |
5ee5090 | Sheng Zha | 07 November 2019, 06:58:30 UTC | Update prepare_clean_env.sh (#423) | 07 November 2019, 06:58:30 UTC |
df20e54 | Jasper Schulz | 05 November 2019, 23:21:11 UTC | Fix edge case in evaluation where prediction length is 1 and prediction target is nan. (#422) | 05 November 2019, 23:21:11 UTC |
481fc5f | Syama Sundar Rangapuram | 04 November 2019, 15:45:08 UTC | DeepStateEstimator: Make cardinality argument compulsory (#413) * DeepStateEstimator: Make cardinality argument compulsory * change test name * addressed comments | 04 November 2019, 15:45:08 UTC |
8642509 | FadhelA | 30 October 2019, 15:13:57 UTC | Dirichlet distribution (#417) * initial commit for dirichlet * fix tests for dirichlet distribution * style fix | 30 October 2019, 15:13:57 UTC |
197866b | Lorenzo Stella | 29 October 2019, 10:39:28 UTC | add question template (#416) * add question template * Update .github/ISSUE_TEMPLATE/question.md Co-Authored-By: Jasper Schulz <jasper.b.schulz@googlemail.com> * align black version/options in github actions to setup.py | 29 October 2019, 10:39:28 UTC |
cd92fd3 | Syama Sundar Rangapuram | 28 October 2019, 23:35:03 UTC | Bug fix: assert all elements in the list are True. (#414) | 28 October 2019, 23:35:03 UTC |
bb70167 | Jasper Schulz | 28 October 2019, 14:43:47 UTC | Remove dataset.validate. (#412) I think this was never used. | 28 October 2019, 14:43:47 UTC |
5713e51 | Syama Sundar Rangapuram | 27 October 2019, 20:17:37 UTC | Use static categorical feature in benchmark_m4 (#410) | 27 October 2019, 20:17:37 UTC |
9c7b966 | Syama Sundar Rangapuram | 25 October 2019, 22:39:07 UTC | Datatype (#363) * Add support for changing dataytpe in DeepAREstimator * Make dtype terminology consistent * undo unnecessary changes added for supporting dtype * need new mxnet version for backward to succeed * fix float_type in backtest * pass dtype explicitly to sample function * Update smoke-test to include tests for dtype | 25 October 2019, 22:39:07 UTC |
b65807d | FadhelA | 24 October 2019, 12:37:36 UTC | bug multivariate gaussian (#407) * bug multivariate gaussian * add test for output shape of low rank multivariate gaussian | 24 October 2019, 12:37:36 UTC |
57746ef | Bo Chang | 23 October 2019, 21:20:09 UTC | Fix empty time_features issue (#387) * Fix empty time_featurs issue * Formatting * Add unit test * Update test/test_transform.py Co-Authored-By: Lorenzo Stella <lorenzostella@gmail.com> | 23 October 2019, 21:20:09 UTC |
4226aa3 | Syama Rangapuram | 23 October 2019, 08:37:08 UTC | Fix batch shape in Binned Distribution | 23 October 2019, 09:46:20 UTC |
93663f0 | Bo Chang | 23 October 2019, 09:20:52 UTC | Update rpy2 requirment in docstring of RForecastPredictor (#405) | 23 October 2019, 09:20:52 UTC |
a894aee | Anton Kiselev | 21 October 2019, 13:00:20 UTC | Dockerfile for GPU container. Fix for installing GPU version of MXNet (#403) * Dockerfile for GPU container. Fix for installing GPU version of MXNet * Typo fix. Replacing requirement without regex. | 21 October 2019, 13:00:20 UTC |
f1f971b | Jasper Schulz | 21 October 2019, 11:00:24 UTC | Batch transform debug (#404) * Added debug to batch-transform. | 21 October 2019, 11:00:24 UTC |
ac782de | Lorenzo Stella | 17 October 2019, 09:27:09 UTC | Add slice_axis methods to Distribution (#397) * add slice, slice_axis to Distribution, add tests * remove slice method * add header * address comment | 17 October 2019, 09:27:09 UTC |
3558cdb | tim | 16 October 2019, 12:16:21 UTC | add more references in README for related scientific papers (#389) * add more references in README for related scientific papers * address comments by jaheba * fix headings hierachy * added a comment on the REFERENCES * including hyperlinks to the papers and clean-up bibtex | 16 October 2019, 12:16:21 UTC |
59912fb | Jasper Schulz | 16 October 2019, 12:08:02 UTC | Stopped CI from posting doc-build links to PRs. (#391) | 16 October 2019, 12:08:02 UTC |
3663066 | Lorenzo Stella | 16 October 2019, 08:56:37 UTC | Fix dim and copy_dim methods in SampleForecast (#366) * fix copy_dim method, adjust docstrings * update tests accordingly * Raise tolerance in sampling test * simplify test code | 16 October 2019, 08:56:37 UTC |
618d4af | Valentin Flunkert | 14 October 2019, 08:36:58 UTC | Increase number of samples for testing distributions (#390) * Increase number of samples for testing distributions * remove flaky annotation * Update test_distribution_sampling.py Removed import. | 14 October 2019, 08:36:58 UTC |
db91331 | tim | 12 October 2019, 22:38:34 UTC | address issue #278 (#388) * address issue #278 * address comments * make sure code formatting is on-board | 12 October 2019, 22:38:34 UTC |
3b47671 | Jasper Schulz | 12 October 2019, 21:59:13 UTC | Simplified dev-setup. (#385) | 12 October 2019, 21:59:13 UTC |
a63d263 | timoschowski | 11 October 2019, 14:32:29 UTC | Fixes #386 Log the number of network parameters after the first forward pass . | 11 October 2019, 14:32:29 UTC |
3095c56 | Jasper Schulz | 10 October 2019, 12:25:42 UTC | Added flaky to test_sampling. (#380) | 10 October 2019, 12:25:42 UTC |
93ebd61 | Jan Gasthaus | 09 October 2019, 13:37:26 UTC | Some small fixes to recipe package (#375) | 09 October 2019, 13:37:26 UTC |
faa8eb4 | Valentin Flunkert | 09 October 2019, 13:24:12 UTC | some small fixes (#377) | 09 October 2019, 13:24:12 UTC |
edcf39b | Lorenzo Stella | 08 October 2019, 17:44:23 UTC | fix flaky sampling test (#373) * fix flaky sampling test * removing useless import | 08 October 2019, 17:44:23 UTC |
0615a80 | Anton Kiselev | 08 October 2019, 14:25:12 UTC | Fixes for serializing sets and numpy numbers in SerDe (#368) * Fixes for serializing sets and numpy numbers in SerDe * Tests for changes in SerDe. Fixes for JSON and code serializations | 08 October 2019, 14:25:12 UTC |
44d7b68 | Lucas Melin | 08 October 2019, 11:58:23 UTC | docs: fix typos in basic and extended forecasting tutorials (#370) | 08 October 2019, 11:58:23 UTC |
cea570d | guilherme-guilherme | 08 October 2019, 11:31:37 UTC | Embedding Input Vector (#315) * Option for DeepAR and DeepState to allow an embedding vector instead of the same value for all categorical features. By default, if no argument, automatically use the heuristic given by Jermey Howard of fast.Ai [1] Each categorical features with size cat_sz, will be associated with a vector with length defined as below. emb_szs = [(c, min(50, (c+1)//2)) for _, c in cat_sz] By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice. [1] https://github.com/fastai/fastai/blob/master/courses/dl1/lesson3-rossman.ipynb * https://github.com/awslabs/gluon-ts/pull/315 Ran "python setup.py type_check" - passed Ran "black src" for proper formatting * Ensuring embedding dimension is not None | 08 October 2019, 11:31:37 UTC |
40c20e3 | Lorenzo Stella | 08 October 2019, 11:12:07 UTC | add recursive batching/unbatching for list and tuple fields (#371) | 08 October 2019, 11:12:07 UTC |
8fb7e33 | Valentin Flunkert | 08 October 2019, 10:54:53 UTC | Revert "Started adding code-owners. (#351)" (#372) This reverts commit ab0befb3503cc476864a98337ff59184290214b2. | 08 October 2019, 10:54:53 UTC |
059e34f | Anton Kiselev | 03 October 2019, 12:27:52 UTC | Using custom SerDe methods for deserializing params in Sagemaker (#364) | 03 October 2019, 12:27:52 UTC |
af91b83 | Bo Chang | 01 October 2019, 21:28:31 UTC | Fix documentation (#362) | 01 October 2019, 21:28:31 UTC |
7e760a4 | Valentin Flunkert | 01 October 2019, 20:12:23 UTC | small fixes (#361) | 01 October 2019, 20:12:23 UTC |