index
:
tensorflow
master
machine learning framework
about
summary
refs
log
tree
commit
diff
homepage
log msg
author
committer
range
path:
root
/
tensorflow
/
python
/
eager
Commit message (
Expand
)
Author
Age
...
*
Reworks the imperative_grad interface.
Alexandre Passos
2017-10-17
*
Better error message for eager-specific APIs
A. Unique TensorFlower
2017-10-16
*
Implement set_shape for EagerTensors for compatibiity with ops that call it
Akshay Agrawal
2017-10-16
*
make_vjp in eager
Alexandre Passos
2017-10-16
*
Proper use of convert_to_tensor in custom_gradient
Alexandre Passos
2017-10-16
*
imperative_grad takes the tape instead of popping it.
Alexandre Passos
2017-10-13
*
Ports the eager gradient tape to C.
Alexandre Passos
2017-10-13
*
eager: Fix issue with custom_gradients and implicit_gradients.
Asim Shankar
2017-10-13
*
Watch trainable variables in graph_callable decorated functions
Akshay Agrawal
2017-10-13
*
eager: Fix an issue with tf.identity.
Asim Shankar
2017-10-10
*
Improve Eager mode random numbers.
A. Unique TensorFlower
2017-10-10
*
Cache last zero tensor in eager gradient computation
Igor Ganichev
2017-10-09
*
TFE: Fix reference counts when copying to Numpy arrays.
Allen Lavoie
2017-10-09
*
New CUDA kernel for LSTMBlockCell's forward propagation.
A. Unique TensorFlower
2017-10-09
*
TFE: Fix reference counts when copying to Numpy arrays.
Allen Lavoie
2017-10-09
*
Removing side outputs from tape code.
Alexandre Passos
2017-10-09
*
Validate input shapes for the graph_callable decorator
Akshay Agrawal
2017-10-09
*
eager: Compute num_gpus() correctly.
Asim Shankar
2017-10-06
*
Convert shape to TensorShape when creating _VariableFromResource
Akshay Agrawal
2017-10-06
*
Fuse TFE_NewOp and TFE_OpGetAttrType to avoid leaking memory.
Allen Lavoie
2017-10-06
*
Start of work towards ordering access to resources in tfe.defun/graph_callable.
Alexandre Passos
2017-10-05
*
eager: Release Python GIL when executing kernels.
Asim Shankar
2017-10-05
*
Add srcs_version="PY2AND3"
Allen Lavoie
2017-10-05
*
Splits backprop.py in two files, one of which can be converted to C
Alexandre Passos
2017-10-05
*
imperative_gradient doesn't fail if some variables are not connected to the o...
Alexandre Passos
2017-10-05
*
Removes use of _grad_fn_accepts_none_for_indices in magic_gradient_function.
Alexandre Passos
2017-10-05
*
Make graph_callable compatible with functions that do not return anything
Akshay Agrawal
2017-10-04
*
Internal cleanup
A. Unique TensorFlower
2017-10-04
*
Added get variable utils to tf.estimator.Estimator.
Mustafa Ispir
2017-10-04
*
Internal cleanup
A. Unique TensorFlower
2017-10-04
*
Make graph_callable, defun tf_decorators
Akshay Agrawal
2017-10-03
*
Removes unnecessary eager-mode call to convert_to_tensor in record_gradient.
Alexandre Passos
2017-10-03
*
Preserve target function signature in custom_gradient decorator
Akshay Agrawal
2017-10-03
*
Internal cleanup.
A. Unique TensorFlower
2017-10-03
*
get_variable in graph_callable returns Variable objects and not Tensors.
Alexandre Passos
2017-10-03
*
Makes custom_gradient work in graph mode.
Alexandre Passos
2017-10-03
*
TFE: Fix tf.layers.Flatten
Shanqing Cai
2017-10-02
*
Add arg name to "op does not support eager execution" error.
A. Unique TensorFlower
2017-10-02
*
Move EagerTensor from python to C.
A. Unique TensorFlower
2017-09-30
*
Merge changes from github.
A. Unique TensorFlower
2017-09-28
*
eager: Remove tfe.device, tf.device suffices.
Asim Shankar
2017-09-28
*
Optimize eager PTB memory to be similar to graph one
A. Unique TensorFlower
2017-09-26
*
Add libtensorflow_framework.so to a bunch of genrules
Allen Lavoie
2017-09-22
*
eager: Fix bug in tfe.gradients_function().
Asim Shankar
2017-09-21
*
Remove RTLD_GLOBAL when loading pywrap_tensorflow
Allen Lavoie
2017-09-21
*
eager: Bugfix for tfe.gradients_function() involving the "Pow" operator.
Asim Shankar
2017-09-21
*
TFE: Simplify code example in backprop.py
Shanqing Cai
2017-09-19
*
Pass around the eager context object instead of getting it
A. Unique TensorFlower
2017-09-19
*
TF Eager: Avoid creating some unnecessary zeros during backprop.
A. Unique TensorFlower
2017-09-18
*
TFE: Add default value back to params of value_and_gradients_function
Shanqing Cai
2017-09-18
[prev]
[next]