Commit history

1. Commits on Jun 13, 2018

   • Remove memory leak in read variable call, and record gradient call.

     Show description for ec769c7
     

     Akshay Modi

     committed
     ec769c7View commit details

     Copy full SHA for ec769c7

     View code at this pointBrowse repository at this point

   • Don't call back into python during insert (which will leave the set in a broken condition if the runtime decides to let another thread run).

     Show description for 50ba6dd
     

     Akshay Modi

     committed
     50ba6ddView commit details

     Copy full SHA for 50ba6dd

     View code at this pointBrowse repository at this point
2. Commits on May 31, 2018

   • Add a subclassed Model's attribute-assigned variables to Model.weights et al

     Show description for 5be69b0
     

     allenlavoie

     authored and

     tensorflower-gardener

     committed
     5be69b0View commit details

     Copy full SHA for 5be69b0

     View code at this pointBrowse repository at this point
3. Commits on May 28, 2018

   • Sort variables in C++ instead of Python.

     Show description for 2d83e13
     

     tomhennigan

     authored and

     tensorflower-gardener

     committed
     2d83e13View commit details

     Copy full SHA for 2d83e13

     View code at this pointBrowse repository at this point
4. Commits on May 25, 2018

   • Ignore while loops instead of mangling them in the automatic control dependencies.

     Show description for e19edb4
     

     alextp

     authored and

     tensorflower-gardener

     committed
     e19edb4View commit details

     Copy full SHA for e19edb4

     View code at this pointBrowse repository at this point

   • Add EagerTensor profiler and device shape utilities

     Show description for ea72dd6
     

     iganichev

     authored and

     tensorflower-gardener

     committed
     ea72dd6View commit details

     Copy full SHA for ea72dd6

     View code at this pointBrowse repository at this point
5. Commits on May 24, 2018

   • When converting a numpy float64 to an EagerTensor, always ensure that it

     Show description for 9532bbb
     

     Akshay Modi

     authored and

     tensorflower-gardener

     committed
     9532bbbView commit details

     Copy full SHA for 9532bbb

     View code at this pointBrowse repository at this point

   • Remove _get_backward_fn and depend on _gradient_function directly.

     Show description for c0dd400
     

     Akshay Modi

     authored and

     tensorflower-gardener

     committed
     c0dd400View commit details

     Copy full SHA for c0dd400

     View code at this pointBrowse repository at this point

   • Fix bugs with the code blocks in defun's docstring.

     Show description for 51645f1
     

     akshayka

     authored and

     tensorflower-gardener

     committed
     51645f1View commit details

     Copy full SHA for 51645f1

     View code at this pointBrowse repository at this point

   • Internal cleanup to remove a difference from the code on github.

     Show description for 437015d
     tensorflower-gardener

     committed
     437015dView commit details

     Copy full SHA for 437015d

     View code at this pointBrowse repository at this point
6. Commits on May 22, 2018

   • Fix memory leak when going from the fast path to the slow path in eager

     Show description for bf6644f
     

     Akshay Modi

     authored and

     tensorflower-gardener

     committed
     bf6644fView commit details

     Copy full SHA for bf6644f

     View code at this pointBrowse repository at this point

   • s/tfe.GradientTape/tf.GradientTape/

     Show description for 45e4c1c
     

     asimshankar

     authored and

     tensorflower-gardener

     committed
     45e4c1cView commit details

     Copy full SHA for 45e4c1c

     View code at this pointBrowse repository at this point
7. Commits on May 21, 2018

   • Fixes issue with gradient tape when asking for the gradient of an intermediate tensor.

     Show description for 753cc5b
     

     alextp

     authored and

     tensorflower-gardener

     committed
     753cc5bView commit details

     Copy full SHA for 753cc5b

     View code at this pointBrowse repository at this point
8. Commits on May 17, 2018

   • Misc changes found when debugging an unrelated problem.

     Show description for 56ba844
     

     Akshay Modi

     authored and

     tensorflower-gardener

     committed
     56ba844View commit details

     Copy full SHA for 56ba844

     View code at this pointBrowse repository at this point

   • Test some distributions stuff in Eager as well as Graph

     Show description for 01dbc6a
     

     Akshay Modi

     authored and

     tensorflower-gardener

     committed
     01dbc6aView commit details

     Copy full SHA for 01dbc6a

     View code at this pointBrowse repository at this point

   • Rename private push/pop API and use from `stop_recording` method.

     Show description for 09a5f58
     

     tomhennigan

     authored and

     tensorflower-gardener

     committed
     09a5f58View commit details

     Copy full SHA for 09a5f58

     View code at this pointBrowse repository at this point

   • Methods to stop and reset tf.GradientTape()

     Show description for 9a75743
     

     alextp

     authored and

     tensorflower-gardener

     committed
     9a75743View commit details

     Copy full SHA for 9a75743

     View code at this pointBrowse repository at this point
9. Commits on May 16, 2018

   • Add performance notes for in-context gradient calls.

     Show description for 8cfbc0c
     

     tomhennigan

     authored and

     tensorflower-gardener

     committed
     8cfbc0cView commit details

     Copy full SHA for 8cfbc0c

     View code at this pointBrowse repository at this point
10. Commits on May 15, 2018

    • Improve documentation for tf.contrib.eager.defun.

      Show description for bb3ebda
      

      akshayka

      authored and

      tensorflower-gardener

      committed
      bb3ebdaView commit details

      Copy full SHA for bb3ebda

      View code at this pointBrowse repository at this point
11. Commits on May 10, 2018

    • Do not differentiate integers in the eager backprop API.

      Show description for 8a8dddf
      

      alextp

      authored and

      tensorflower-gardener

      committed
      8a8dddfView commit details

      Copy full SHA for 8a8dddf

      View code at this pointBrowse repository at this point

    • Automated g4 rollback of changelist 195878952

      Show description for e696dc1
      

      asimshankar

      authored and

      tensorflower-gardener

      committed
      e696dc1View commit details

      Copy full SHA for e696dc1

      View code at this pointBrowse repository at this point
12. Commits on May 9, 2018

    • Delete old op gen code and replace with eager op gen.

      Show description for 7bd992b
      

      Akshay Modi

      authored and

      tensorflower-gardener

      committed
      7bd992bView commit details

      Copy full SHA for 7bd992b

      View code at this pointBrowse repository at this point

    • Include tensorflow::DataType header file

      Show description for ffe6ede
      

      Akshay Modi

      authored and

      tensorflower-gardener

      committed
      ffe6edeView commit details

      Copy full SHA for ffe6ede

      View code at this pointBrowse repository at this point

    • Make eager functions runable on TPU

      Show description for 14d5f21
      

      iganichev

      authored and

      tensorflower-gardener

      committed
      14d5f21View commit details

      Copy full SHA for 14d5f21

      View code at this pointBrowse repository at this point
13. Commits on May 8, 2018

    • Do not differentiage integers in the eager API.

      Show description for f58effe
      

      alextp

      authored and

      tensorflower-gardener

      committed
      f58effeView commit details

      Copy full SHA for f58effe

      View code at this pointBrowse repository at this point

    • When building functions, capture tensors in `internal_convert_to_tensor`.

      Show description for 83aa323
      

      akshayka

      authored and

      tensorflower-gardener

      committed
      83aa323View commit details

      Copy full SHA for 83aa323

      View code at this pointBrowse repository at this point

    • Automated g4 rollback of changelist 195748721

      Show description for a799cdb
      tensorflower-gardener

      committed
      a799cdbView commit details

      Copy full SHA for a799cdb

      View code at this pointBrowse repository at this point

    • Make eager functions runable on TPU

      Show description for b6906d1
      

      iganichev

      authored and

      tensorflower-gardener

      committed
      b6906d1View commit details

      Copy full SHA for b6906d1

      View code at this pointBrowse repository at this point
14. Commits on Apr 30, 2018

    • Fix typos in tf.GradientTape documentation.

      Show description for 6e9d8ab
      

      tomhennigan

      authored and

      tensorflower-gardener

      committed
      6e9d8abView commit details

      Copy full SHA for 6e9d8ab

      View code at this pointBrowse repository at this point

    • Clarify return type for defun as zero or more `tf.Tensor`s.

      Show description for 1872f29
      

      tomhennigan

      authored and

      tensorflower-gardener

      committed
      1872f29View commit details

      Copy full SHA for 1872f29

      View code at this pointBrowse repository at this point

    • Fixes to tape gradient for providing outputs and having multiple targets.

      Show description for aa2405e
      

      alextp

      authored and

      tensorflower-gardener

      committed
      aa2405eView commit details

      Copy full SHA for aa2405e

      View code at this pointBrowse repository at this point
15. Commits on Apr 27, 2018

    • eager: Improve error message when GradientTape is used incorrectly.

      Show description for de26d3e
      

      asimshankar

      authored and

      tensorflower-gardener

      committed
      de26d3eView commit details

      Copy full SHA for de26d3e

      View code at this pointBrowse repository at this point
16. Commits on Apr 26, 2018

    • For tf.gradients(), do not backpropagate through integer tensors.

      Show description for f637506
      tensorflower-gardener

      committed
      f637506View commit details

      Copy full SHA for f637506

      View code at this pointBrowse repository at this point

    • Simplify tfe.defun capture by not using convert_to_tensor

      Show description for 509ffc3
      

      alextp

      authored and

      tensorflower-gardener

      committed
      509ffc3View commit details

      Copy full SHA for 509ffc3

      View code at this pointBrowse repository at this point
17. Commits on Apr 24, 2018

    • Make TF functions work with _USE_C_SHAPES=True.

      Show description for e7db82f
      

      skye

      authored and

      tensorflower-gardener

      committed
      e7db82fView commit details

      Copy full SHA for e7db82f

      View code at this pointBrowse repository at this point