Commit message (Collapse) | Author | Age | |
---|---|---|---|
* | Updating the V2 variables API. | Alexandre Passos | 2018-09-27 |
| | | | | PiperOrigin-RevId: 214824023 | ||
* | Move from deprecated self.test_session() to self.cached_session(). | A. Unique TensorFlower | 2018-09-21 |
| | | | | | | | | self.test_session() has been deprecated in 9962eb5e84b15e309410071b06c2ed2d6148ed44 as its name confuses readers of the test. Moving to cached_session() instead which is more explicit about: * the fact that the session may be reused. * the session is not closed even when doing a "with self.test_session()" statement. PiperOrigin-RevId: 213944355 | ||
* | address lint again | Peng Yu | 2018-05-21 |
| | |||
* | address lint | Peng Yu | 2018-05-21 |
| | |||
* | address comments | Peng Yu | 2018-05-21 |
| | |||
* | add test case always predict [0, 1] | Peng Yu | 2018-05-21 |
| | |||
* | add inference support for tree and forest variables | Peng Yu | 2018-05-21 |
| | |||
* | Fix warnings in tf.contrib.tensor_forest | Yong Tang | 2018-04-22 |
| | | | | Signed-off-by: Yong Tang <yong.tang.github@outlook.com> | ||
* | contrib: minor spelling tweaks | Brett Koonce | 2018-04-07 |
| | | | | | | | packages: data training tensor_forest | ||
* | Output variance over tree predictions for classifications. | A. Unique TensorFlower | 2018-01-05 |
| | | | | PiperOrigin-RevId: 180976319 | ||
* | Clarify expectations about the input_data parameter. | A. Unique TensorFlower | 2017-10-02 |
| | | | | PiperOrigin-RevId: 170745215 | ||
* | Add README.md for TensorForest. | A. Unique TensorFlower | 2017-09-14 |
| | | | | PiperOrigin-RevId: 168685581 | ||
* | Make sure variance doesn't go below zero from floating point errors in ↵ | A. Unique TensorFlower | 2017-08-02 |
| | | | | | | TensorForest. PiperOrigin-RevId: 164033338 | ||
* | Add regression variance over individual trees to TensorForest inference. | A. Unique TensorFlower | 2017-07-31 |
| | | | | PiperOrigin-RevId: 163695881 | ||
* | Optionally output a new TreePath proto during TensorForest inference for ↵ | A. Unique TensorFlower | 2017-07-28 |
| | | | | | | ultimate interpretability. PiperOrigin-RevId: 163466324 | ||
* | Fix TensorForest's saveable object names so loading a savedmodel works. | A. Unique TensorFlower | 2017-07-27 |
| | | | | PiperOrigin-RevId: 163332598 | ||
* | Remove old TensorForest implementation. | A. Unique TensorFlower | 2017-07-24 |
| | | | | PiperOrigin-RevId: 162944683 | ||
* | Destroy resources at end of session in TensorForestEstimator to avoid memory ↵ | A. Unique TensorFlower | 2017-07-18 |
| | | | | | | leak. PiperOrigin-RevId: 162383623 | ||
* | Automated g4 rollback of changelist 162273792 | Anna R | 2017-07-17 |
| | | | | PiperOrigin-RevId: 162296574 | ||
* | TensorForest v4 kernel tests for proto-based model. Also some stuff to ↵ | A. Unique TensorFlower | 2017-07-17 |
| | | | | | | include decision_trees in pip package properly. PiperOrigin-RevId: 162273792 | ||
* | Switch TensorForestEstimator to tf.learn heads lib. | A. Unique TensorFlower | 2017-07-14 |
| | | | | PiperOrigin-RevId: 161949540 | ||
* | Refactor some of TensorForest V4 to make the tree model valid during ↵ | A. Unique TensorFlower | 2017-07-12 |
| | | | | | | training time, instead of only after FinalizeTreeOp. PiperOrigin-RevId: 161663317 | ||
* | Add option to TensorForest to initialize splits with the average of 2 examples. | A. Unique TensorFlower | 2017-07-06 |
| | | | | PiperOrigin-RevId: 161122345 | ||
* | Improve data processing in TensorForest. Extract name correctly when ↵ | A. Unique TensorFlower | 2017-06-28 |
| | | | | | | transforming, and support sparse categorical. PiperOrigin-RevId: 160424501 | ||
* | Migrate TensorForest v4 python to contrib. | A. Unique TensorFlower | 2017-06-27 |
| | | | | PiperOrigin-RevId: 160308805 | ||
* | Migrate ops for new version of TensorForest. | A. Unique TensorFlower | 2017-06-22 |
| | | | | PiperOrigin-RevId: 159852889 | ||
* | Automated g4 rollback of changelist 159718610 | A. Unique TensorFlower | 2017-06-21 |
| | | | | PiperOrigin-RevId: 159728380 | ||
* | Migrate ops for new version of TensorForest. | A. Unique TensorFlower | 2017-06-21 |
| | | | | PiperOrigin-RevId: 159718610 | ||
* | Fix TensorForest feature bagging when using feature-column data. | A. Unique TensorFlower | 2017-06-20 |
| | | | | PiperOrigin-RevId: 159566463 | ||
* | Automated g4 rollback of changelist 152936284 | A. Unique TensorFlower | 2017-05-10 |
| | | | | PiperOrigin-RevId: 155652925 | ||
* | Reshape all dense features to a rank 2, shape (*,1) tensor to avoid concat | A. Unique TensorFlower | 2017-04-12 |
| | | | | | errors. Change: 152936284 | ||
* | Fix the dlopen contrib test hack by making a pywrap_tensorflow module that ↵ | Jonathan Hseu | 2017-02-28 |
| | | | | | | | | | imports pywrap_tensorflow_internal with RTLD_GLOBAL. Fixes #6568 Change: 148843302 | ||
* | Make TensorForest dense input shape a little more flexible by inferring size ↵ | A. Unique TensorFlower | 2017-02-28 |
| | | | | | | 1 if needed. Change: 148825114 | ||
* | Add ability to report feature importances in TensorForest. | A. Unique TensorFlower | 2017-02-23 |
| | | | | Change: 148383964 | ||
* | Make TensorForest use a more idiomatic device assigner ↵ | A. Unique TensorFlower | 2017-02-21 |
| | | | | | | (VariableDeviceChooser by default), specifying simply that variables/ops for trees should be grouped. Change: 148141071 | ||
* | Remove unused topn code. | A. Unique TensorFlower | 2017-02-18 |
| | | | | Change: 147940756 | ||
* | Make random_forest_test TSAN-compliant by inserting some dependencies to ↵ | A. Unique TensorFlower | 2017-01-31 |
| | | | | | | avoid R/W hazards. Change: 146121173 | ||
* | Switch tf.concat_v2 references in third_party/tensorflow to tf.concat. | A. Unique TensorFlower | 2017-01-10 |
| | | | | Change: 144153795 | ||
* | Support multiple trainers in TensorForest that split trees among them. | A. Unique TensorFlower | 2017-01-10 |
| | | | | Change: 144086694 | ||
* | Create a class to use for input specification in TensorForest, and support a ↵ | A. Unique TensorFlower | 2017-01-09 |
| | | | | | | | | mix of dense and sparse features. Move feature processing to tensor_forest to have the most sane interface (just dict of tensors or single tensor). Had to move data_ops because of python3 import weirdness. Change value == bias to go right instead of left because this inherently handles sparse one-hot categorical data (if a node's bias is set at 1 (the only value it could pick), both values of 1 and 0 are not strictly > than 1, so that node would always go left). Change: 143970989 | ||
* | Remove so many more hourglass imports | Justine Tunney | 2016-12-29 |
| | | | | Change: 143230429 | ||
* | Rename usages of tf.mul, tf.neg, tf.sub that are used internally | Andrew Selle | 2016-12-20 |
| | | | | Change: 142595367 | ||
* | Switch array_ops.pack/unpack to array_ops.stack/unstack. Also switch a few ↵ | A. Unique TensorFlower | 2016-12-15 |
| | | | | | | remaining references to tf.pack/unpack to tf.stack/unstack. Change: 142108785 | ||
* | Major intent of this CL is to rename split_v -> split in the python API. | A. Unique TensorFlower | 2016-12-12 |
| | | | | | | | | | | | | | Requires: 1) Add name arguments to tf.split calls introduced since major Rosie CL cleaning this up across the codebase. 2) Change uses of array_ops.split to use named arguments, which was not covered in the Rosie CL. 3) Rename split_v calls to split. Change: 141806936 | ||
* | Merge TensorForest ops into one library/generated python file, getting rid ↵ | A. Unique TensorFlower | 2016-12-12 |
| | | | | | | | of useless template code. Load ops at import time, which is more idiomatic with the rest of tf/contrib. Also rename TensorForests string_to_float op to be more descriptive and unique (reinterpret_string_to_float). Change: 141756823 | ||
* | SparseTensor.shape -> SparseTensor.dense_shape part 2... | Eugene Brevdo | 2016-12-08 |
| | | | | Change: 141460319 | ||
* | Switch all tf.concat(concat_dim, value, name) calls in ↵ | A. Unique TensorFlower | 2016-12-06 |
| | | | | | | third_party/tensorflow to tf.concat_v2(value, axis, name). Change: 141255675 | ||
* | Force fix documentation causing open source blocks. | Andrew Selle | 2016-12-06 |
| | | | | Change: 141201358 | ||
* | removing tf.slice as it is soon to be deprecated. replaced with ↵ | Olivia Nordquist | 2016-12-02 |
| | | | | | | tf.strided_slice in the following 15 files. more to come! Change: 140875384 | ||
* | Change .fill() formula for num_splits_to_consider. | A. Unique TensorFlower | 2016-11-30 |
| | | | | Change: 140684634 |