[go: nahoru, domu]

Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Branch 152141388 #8958

Merged
merged 82 commits into from
Apr 4, 2017
Merged

Branch 152141388 #8958

merged 82 commits into from
Apr 4, 2017

Conversation

rohan100jain
Copy link
Member

No description provided.

tensorflower-gardener and others added 30 commits March 31, 2017 01:34
…ell instance argument support.

Change: 151829043
… folding

pass is applied before LayoutOptimizer.
Change: 151830580
Previously Clamp(f32[5], f32[], f32[9]) returned success, but it now returns a
failure.  Noticed while debugging a different problem.
Change: 151835981
We only need to worry about overflowing uint64, not int64.
Change: 151836590
…er classes tensor.

Servo expects classes to be a string tensor of the same shape as scores and containing the labels for corresponding scores. While creating output alternatives, if classes tensor does not match these conditions, we create a new tensor with these properties.
Change: 151838225
…:16.04-based gcs_test Dockerfile

Change: 151843345
Compile nccl on windows, now that bazel 0.4.5 is used.
Change: 151853954
Change: 151860614
ZF removed the restriction under the hood, but didn't remove the check.

Github issue for similar tf.tile / binary operator support: #8873.
Change: 151863009
…an array instead of a list of indices.

Change: 151863667
This makes it easier to create a zero_state and insert, say, the encoder's state
directly into it, using public methods.
Change: 151868236
Change: 151876042
tensorflower-gardener and others added 25 commits April 3, 2017 14:07
… support mobile SOC experiments and leverage tf tools

Change: 152066347
Continuing the effort to let people train neural networks
the size of a small galaxy, or get coherent errors when
they can't.
Change: 152079358
…ar string tensors.

If the string tensor passed to the text dashboard is not a scalar, then the data is automatically organized into a table.
Change: 152081823
This will serve as a regression test for future CLs changing how
functions are instantiated in tfcompile.
Change: 152085567
…CoupledInputForgetGateLSTMCell, GridLSTMCell and AttentionCellWrapper.

Change: 152086191
…stDebugTrainingDynamicRNNWorks

Change: 152119214
…onBuilder::ComputeConstant() should not be persistent.

The reason to note failures via first_error_ is so that most methods can defer failure reporting until Build() time. However this makes no sense for methods such as ComputeConstant() that return a Status already -- such methods can just return their errors directly without "polluting" other callers.
Change: 152129004
…This will allow existing custom estimators to easily switch to core estimator until we provide better support (like moving heads to core).

Change: 152133991
@yifeif yifeif merged commit efe5376 into tensorflow:master Apr 4, 2017
copybara-service bot pushed a commit that referenced this pull request Feb 6, 2024
Imported from GitHub PR openxla/xla#8958

Copybara import of the project:

--
c9e5808563b24063dac97aec2585ed6e9112c0b7 by Dragan Mladjenovic <Dragan.Mladjenovic@amd.com>:

[ROCm] Adapt failing hlo tests

Merging this change closes #8958

PiperOrigin-RevId: 604496716
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet