| Commit message (Collapse) | Author | Age |
|
|
|
|
|
|
|
|
|
|
| |
*** Original change description ***
Add HloSchedule class representing a sequential order of an HloModule.
Currently we represent a sequential schedule of a module using a SequentialHloOrdering::HloModuleSequence which is a type alias of a bare map from HloComputation* to std::vector<HloInstruction*>. This CL replaces this with a proper class which results in better encap...
***
PiperOrigin-RevId: 211726890
|
|
|
|
|
|
|
|
| |
TF 2.0 to return a no-arg callable to output a learning rate, instead of directly outputting a learning rate tensor.
This brings the graph mode API in line with the eager execution API, where this change was made to allow changing the learning rate value across different invocations of optimizer functions.
PiperOrigin-RevId: 211726295
|
|
|
|
| |
PiperOrigin-RevId: 211724843
|
|
|
|
| |
PiperOrigin-RevId: 211722113
|
|
|
|
|
|
| |
Also upload fully quantized mobilenet-v2 and inception-v3 models.
PiperOrigin-RevId: 211721504
|
|
|
|
| |
PiperOrigin-RevId: 211721354
|
|
|
|
| |
PiperOrigin-RevId: 211719399
|
|
|
|
| |
PiperOrigin-RevId: 211719342
|
|
|
|
| |
PiperOrigin-RevId: 211719009
|
|
|
|
|
|
| |
RELNOTES: n/a
PiperOrigin-RevId: 211718248
|
|
|
|
|
|
| |
CancellationManagers for each run of the function.
PiperOrigin-RevId: 211717580
|
|
|
|
| |
PiperOrigin-RevId: 211715608
|
|
|
|
| |
PiperOrigin-RevId: 211714574
|
|
|
|
| |
PiperOrigin-RevId: 211712907
|
|
|
|
| |
PiperOrigin-RevId: 211711493
|
|
|
|
|
|
| |
second axis of Tensors through tf.train.warm_start. Note that the underlying initializer already has this functionality (for example, for output layers).
PiperOrigin-RevId: 211709879
|
|
|
|
|
|
| |
These APIs are based on queue runners, which have been deprecated and will be removed in TensorFlow 2.0. They have been replaced with `tf.data.Dataset`, which provides a more efficient version of the same functionality.
PiperOrigin-RevId: 211708268
|
|
|
|
|
|
|
|
|
| |
There's no need for the two separate classes anymore. This also cleans up some other parts of the interface:
* Removes the clear_resource_control_flow_state, which isn't used anywhere
* Makes capture_value a private method of FuncGraph (_capture_helper)
* Makes create_substitute_placeholder private
PiperOrigin-RevId: 211707906
|
|
|
|
| |
PiperOrigin-RevId: 211707155
|
|
|
|
|
|
| |
implementations of the batch and padded batch reducers.
PiperOrigin-RevId: 211706766
|
|\
| |
| |
| | |
PiperOrigin-RevId: 211706322
|
| |
| |
| |
| | |
PiperOrigin-RevId: 211705274
|
|\ \
| | |
| | |
| | | |
PiperOrigin-RevId: 211705018
|
| | |
| | |
| | |
| | |
| | |
| | | |
This still defaults to float32 for all normal floats.
PiperOrigin-RevId: 211704918
|
| | |
| | |
| | |
| | | |
PiperOrigin-RevId: 211703281
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
AccumulateNV2 doesn't have or need a kernel. It gets rewritten to other ops by
accumulate_n_optimizer.cc. This change allows it to be mentioned in the output
of print_selective_registration_header, rather than being ignored with a
warning. Behavior for other ops is preserved.
PiperOrigin-RevId: 211701878
|
| | |
| | |
| | |
| | |
| | |
| | | |
These APIs are based on queue runners, which have been deprecated and will be removed in TensorFlow 2.0. They have been replaced with `tf.data.Dataset`, which provides a more efficient version of the same functionality.
PiperOrigin-RevId: 211700442
|
|\ \ \
| | | |
| | | |
| | | | |
PiperOrigin-RevId: 211700362
|
| | | |
| | | |
| | | |
| | | | |
PiperOrigin-RevId: 211698400
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
The "Proto" suffix adds little clarity but makes a long type name even longer.
PiperOrigin-RevId: 211693871
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
consistently
StringPiece is an alias for absl::string_view, InlinedVector is aliased to absl::InlinedVector. StrCat is compatible, so swapping it out is safe.
PiperOrigin-RevId: 211691840
|
| | | |
| | | |
| | | |
| | | | |
PiperOrigin-RevId: 211688974
|
| | | |
| | | |
| | | |
| | | | |
PiperOrigin-RevId: 211687433
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Automated rollback of commit 7fa693209fe238478739b3982f652a7e35be91f3
PiperOrigin-RevId: 211681957
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
benchmark on O-MR1 without an exit() of the process.
Also fixes bug in interpretation of error values (NNAPI vs. TFLite error
codes).
PiperOrigin-RevId: 211681942
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
I want --vmodule=xla_compilation_cache=1 to print only the most essential
things.
PiperOrigin-RevId: 211676846
|
| | | |
| | | |
| | | |
| | | | |
PiperOrigin-RevId: 211666438
|
| | | |
| | | |
| | | |
| | | | |
PiperOrigin-RevId: 211665268
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This modifies
https://github.com/tensorflow/tensorflow/commit/834da2c3fddab1bbbce742db572cfe65dd320fcd
to work with tfe.defun in addition to the legacy Defun implementation.
PiperOrigin-RevId: 211663702
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
duplicate definitions.
PiperOrigin-RevId: 211662523
|
| | | |
| | | |
| | | |
| | | | |
PiperOrigin-RevId: 211661670
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
use_gpu does not affect the creation of the session, it only affects the
context manager in which nodes are added to the graph, so it should not
be included in the consistency check.
PiperOrigin-RevId: 211659833
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
LSTM Op.
This introduces a connection between forward and backward cells across subsequent layers when stacking bidirectional LSTM Ops on top of each other.
In more detail:
Previously, the Op had only one input that was fed into the layer in the
following way:
INPUT (INPUT_REVERSED)
| |
-----------------------
| FW_LSTM BW_LSTM | <----- bidi-LSTM cell (with one input / two outputs)
-----------------------
| |
FW_OUT BW_OUT
Now, the Op can have an (optional) auxiliary input in the following way:
AUX_INPUT (AUX_INPUT_REVERSED)
| |
INPUT | (INPUT_R'D.)|
| | | |
-------------------------
| \ / \ / |
| FW_LSTM BW_LSTM | <----- bidi-LSMT cell (with 2 inputs / 2 outputs)
-------------------------
| |
FW_OUT BW_OUT
When stacking these Ops, previously, only the following flow was allowed:
Input
/ \
FW_LSTM1 BW_LSTM1
| |
| |
FW_LSTM2 BW_LSTM2
| |
| |
FW_LSTM3 BW_LSTM3
\ /
Output
With the introduction of an auxiliary input to the bidi-LSTM layer, the forward
(FW_LSTMi) output of the ith layer is fed into as the input to the next layer
(hence, inputs to both FW_LSTM{i+1} and BW_LSTM{i+1}) and the backward output is
fed as the auxiliary inputs to both FW_LSTM{i+1} and BW_LSTM{i+1}). This way, the
stacking can be changed to allow for the "cross-linking" between subsequent
layer in the following way:
Input
/ \
FW_LSTM1 BW_LSTM1
| \ / |
| / \ |
FW_LSTM2 BW_LSTM2
| \ / |
| / \ |
FW_LSTM3 BW_LSTM3
\ /
Output
PiperOrigin-RevId: 211659472
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Currently we represent a sequential schedule of a module using a SequentialHloOrdering::HloModuleSequence which is a type alias of a bare map from HloComputation* to std::vector<HloInstruction*>. This CL replaces this with a proper class which results in better encapsulation of code which deals with schedules and better enforcement of invariants.
This CL also fixes a corner-case bug in dataflow analysis, where values of instructions which are live out of the computation erroneously did not interfere with the values of instructions scheduled after the root instruction.
PiperOrigin-RevId: 211656888
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
~25-30% speedup when compiled with AVX.
* collapse inner dims before contraction
* eval kernel tensor before contraction
PiperOrigin-RevId: 211651030
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This prevents these build-time rules from accessing any GPUs which might
be present on the build machine and interfering with GPU tests which
might be running concurrently.
PiperOrigin-RevId: 211647681
|
| | | |
| | | |
| | | |
| | | | |
PiperOrigin-RevId: 211643209
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
accumulator.
PiperOrigin-RevId: 211642436
|
| | | |
| | | |
| | | |
| | | | |
PiperOrigin-RevId: 211639440
|
| | | |
| | | |
| | | |
| | | | |
PiperOrigin-RevId: 211637019
|