Skip to content

Latest commit

 

History

History
39 lines (33 loc) · 3.46 KB

CHANGES.md

File metadata and controls

39 lines (33 loc) · 3.46 KB

Breaking API changes.

tensorpack is still in early development, and API changes can happen. The backward compatibilty will be preserved for several months, with a deprecation warning, so you won't need to look at here very often.

Here are a list of things that were changed, starting from an early version. TensorFlow itself also changes API and those are not listed here.

  • 2017/02/20. The interface of step callbacks are changed to be the same as tf.train.SessionRunHook. If you haven't written any custom step callbacks, there is nothing to do. Otherwise please refer to the existing callbacks.
  • 2017/02/12. TrainConfig(optimizer=) was deprecated. Now optimizer is set in ModelDesc. And gradient processors become part of an optimizer. See commit.
  • 2017/02/11. _get_input_vars() in ModelDesc was renamed to _get_inputs. InputVar was renamed to InputDesc. See commit.
  • 2017/01/27. TrainConfig(step_per_epoch) was renamed to steps_per_epoch. See commit.
  • 2017/01/25. Argument order of models.ConcatWith is changed to follow the API change in TensorFlow upstream. See commit.
  • 2017/01/25. TrainConfig(callbacks=) now takes a list of Callback instances. See commit on how to change your code.
  • 2017/01/06. summary.add_moving_summary now takes any number of positional arguments instead of a list. See commit to change your code.
  • 2017/01/05. The argument TrainConfig(dataset=) is renamed to TrainConfig(dataflow=). See commit to change your code.
  • 2016/12/15. The predict_tower option is in TrainConfig now instead of Trainer. See commit.
  • 2016/11/10. The {input,output}_var_names argument in PredictConfig is renamed to {input,output}_names. See commit.
  • 2016/11/06. The inferencer ClassificationError now expects the vector tensor returned by prediction_incorrect instead of the "wrong" tensor. See commit to make changes.
  • 2016/10/17. Conv2D and FullyConnect use tf.identity by default instead of tf.nn.relu. See commit.
  • 2016/09/01. The method _build_graph of ModelDesc doesn't take is_training argument anymore. The is_training attribute can be obtained from tower context. See commit.
  • 2016/05/15. The method _get_cost of ModelDesc is replaced by _build_graph. See commit.