Create your own GitHub profile
Sign up for your own profile on GitHub, the best place to host code, manage projects, and build software alongside 50 million developers.
Sign up
Popular repositories
-
-
Forked from tensorflow/tensorflow
Computation using data flow graphs for scalable machine learning
C++
-
-
-
1,417 contributions in the last year
Contribution activity
June 2020
Created a pull request in google/jax that received 4 comments
Attach source info to Jaxpr equations.
Example:
In [1]: import jax, jax.numpy as jnp
In [2]: def f(x, y): ...: z = jnp.cos(x) ...: z = z * jnp.tanh(y) ...: return z + 2 ...: In [3]: prin…
+158
−60
•
4
comments
- Add a heap profiler API and document it.
- Update README for jaxlib 0.1.50 release.
- Fix test failures on GPU.
- Update docker script for CUDA 11 and libcudnn8.
- Update XLA in preparation for a new jaxlib release (0.1.50).
- Make CUDA wheels manylinux2010 compliant, add CUDA 11, drop CUDA 9.2
- Fix eigh JVP to ensure that both the primal and tangents of the eigen…
- Github actions jaxlib build experiment.
- Fix failure in PyTorch array interoperability test in non-x64 mode.
- Remove warning suppression for tuple and list arguments to reductions.
- Avoid direct type/dtype comparisons to fix NumPy 1.19 deprecation war…
- Fix test failures due to RuntimeWarning in variance tests.
- Restrict .compress tests to arrays only.
- Merge quantile and nanquantile implementations.
- Add missing functions to documentation.
- Fix test failure in jax2tf due to conflicting merges.
- Bump XLA version.
- Change build badge from Travis to Github actions.
- Switch CI builds from Travis to Github actions
- Minor change to test presubmits.
- Small edit to documentation.
- Fix memory leak when no axis is provided to pmap.
- Disable some masking tests on TPU.
- Fix failures in masking_test.py
- Proofread documentation.
- Some pull requests not shown.
- Fix typos "Pytrees" page on readthedocs
- add remat docstring
- Clarify docstrings regarding usage of static arguments in jit and vmap.
- Fixed issue with the 'compress' method of 'jnp.ndarray'
- Future warning about lists and tuples
- WIP: Print JAX arrays to full precision
- Add np.unwrap
- Add segment_max operator.
- lax.sort: allow any sequence of Arrays, not just tuples
- Adds boolean support for bitwise not and unittests for boolean suppor…
- Add instructions for how to use the TensorBoard profiler to the profiling docs.
- Disable the workaround to prevent expansion of type aliases
- Remove documentation presubmit action now that the RTD presubmit is enabled.
- make jnp.array(x, copy=True) copies device buffers
- Revert "Initial implementation of variadic lax.reduce() (#3342)"
- flake8: change from inclusion list to exclusion list
- fix flakes at head
- Refined the forcing of annotation processig in doc building
- deflake remainder of jax
- Cleanup: deflake jax.experimental and jax.ops
- Extract JAX primitive test harness from lax_test.py.
- deflake jax/lax & add to flake8 check
- Cleanup: de-lint tests directory & add flake8 to travis
- Fix broken jnp.nancumsum() & jnp.nancumprod() and add tests
- Fix validation code in lax.conv
- Some pull request reviews not shown.