| Commit message (Collapse) | Author | Age |
|
|
|
| |
Change: 147596183
|
|
|
|
|
| |
Error introduced during CL/147543432 touchups.
Change: 147594098
|
|
|
|
| |
Change: 147590973
|
|
|
|
| |
Change: 147563588
|
|
|
|
| |
Change: 147563127
|
|
|
|
| |
Change: 147562709
|
|
|
|
| |
Change: 147561890
|
|
|
|
|
|
|
| |
Change failure to meet memory limit during rematerialization from
a fatal error to a WARNING message. Rematerialization only estimates
runtime memory usage and the estimate is not even a bound (lower or upper).
Change: 147559827
|
|
|
|
|
|
| |
RuntimeError: Modules nested too deep:
contrib.graph_editor.reroute.subgraph.select.tf_ops.dtypes.np.core.fromnumeric.mu
Change: 147558652
|
|
|
|
| |
Change: 147558587
|
|
|
|
| |
Change: 147554230
|
|
|
|
|
| |
Tensorflow contrib.
Change: 147549897
|
|
|
|
| |
Change: 147543652
|
|
|
|
|
| |
Export DynamicRnnEstimator as part of tf.contrib.learn.
Change: 147543432
|
|
|
|
|
|
| |
FixedUnigramSampler::Probability is too large.
Change: 147539715
|
|
|
|
| |
Change: 147539687
|
|
|
|
| |
Change: 147538352
|
|
|
|
| |
Change: 147531388
|
|
|
|
|
|
|
| |
Fails: http://ci.tensorflow.org/view/Tensorflow%20Jenkins%20Monitored%20builds/job/tensorflow-master-cpu/1767/console
http://ci.tensorflow.org/view/Tensorflow%20Jenkins%20Monitored%20builds/job/tensorflow-master-cpu/1766/console
Change: 147529313
|
|
|
|
| |
Change: 147527599
|
|
|
|
|
| |
reference, don't also put `...` around it.
Change: 147522473
|
|
|
|
| |
Change: 147519957
|
|
|
|
| |
Change: 147519438
|
|
|
|
|
|
| |
This is in lieu of a more detailed description of how to use Tensor
objects, which we should probably add at some point.
Change: 147517482
|
|
|
|
| |
Change: 147514603
|
|
|
|
| |
Change: 147514084
|
|
|
|
|
| |
numpy 1.12 only.
Change: 147512538
|
|
|
|
|
| |
Also move the embedding projector tests to their own file (projector/plugin_test.py).
Change: 147511031
|
|
|
|
|
| |
argparse. Fix a bool flag and also pass only unparsed flags to the test.main.
Change: 147508611
|
|
|
|
|
|
| |
DynamicUpdateSlice instructions.
Change: 147506969
|
|
|
|
| |
Change: 147499520
|
|
|
|
| |
Change: 147497724
|
|
|
|
|
|
| |
while loops
Change: 147488620
|
|
|
|
| |
Change: 147487962
|
|
|
|
|
| |
Use the C++11 version of std::fe{set,get}round defined in <cfenv>.
Change: 147487955
|
|
|
|
| |
Change: 147487768
|
|
|
|
|
|
|
| |
for easier building of the demo's native code in a standalone manner.
libtensorflow_inference.so is now added to the demo app independently for TF support.
Change: 147485685
|
|
|
|
|
|
| |
inheritance.
Change: 147480076
|
|
|
|
|
|
|
| |
* Skip @@ lines in docstrings.
* Do @{...} conversion in brief docstrings.
* Fix a doc linking bug.
Change: 147473081
|
|
|
|
| |
Change: 147473069
|
|
|
|
| |
Change: 147470630
|
|
|
|
| |
Change: 147452567
|
|
|
|
| |
Change: 147441750
|
|
|
|
| |
Change: 147437030
|
|
|
|
|
|
|
|
|
| |
tf.contrib.seq2seq.
1. Expose the new classes and functions in tf.contrib.seq2seq
2. Delete old functions
3. Rename dynamic_decode_rnn -> dynamic_decode
Change: 147436900
|
|
|
|
|
| |
TASK 2. //training/... Python class doc strings
Change: 147434408
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Old:
/// Copy Op.
///
/// Performs CPU-to-CPU or GPU-to-GPU deep-copying of tensor, depending on the
/// device on which the tensor is allocated.
///
/// Unlike the CopyHost Op, this op does not have HostMemory constraint on its
/// input or output.
///
/// Arguments:
/// * scope: A Scope object
/// * input: Input tensor.
///
/// Returns:
/// * `Output`: Output tensor, deep-copied from input.
class Copy {...}
New:
/// Copy Op.
///
/// Performs CPU-to-CPU or GPU-to-GPU deep-copying of tensor, depending on the
/// device on which the tensor is allocated.
///
/// Unlike the CopyHost Op, this op does not have HostMemory constraint on its
/// input or output.
///
/// Arguments:
/// * scope: A Scope object
/// * input: Input tensor.
///
/// Optional attributes (see `Attrs`):
/// * tensor_name: The name of the input tensor.
///
/// Returns:
/// * `Output`: Output tensor, deep-copied from input.
class Copy {...}
Change: 147432712
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Old:
/// Optional attribute setters for Copy :
///
/// TensorName(StringPiece): Defaults to ""
/// The name of the input tensor.
struct Attrs {
Attrs TensorName(StringPiece x) {
Attrs ret = *this;
ret.tensor_name_ = x;
return ret;
}
StringPiece tensor_name_ = "";
};
New:
/// Optional attribute setters for Copy
struct Attrs {
/// The name of the input tensor.
///
/// Defaults to ""
Attrs TensorName(StringPiece x) {
Attrs ret = *this;
ret.tensor_name_ = x;
return ret;
}
StringPiece tensor_name_ = "";
};
Change: 147430152
|
|
|
|
|
|
| |
-Seal and expose reroute
-Remove unavailable aliases from guide
Change: 147429891
|
|
|
|
| |
Change: 147428303
|