experiments/STT-tensorflow
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Add release note.

Browse Source
This commit is contained in:
Nathan Luehr 2021-01-13 12:51:04 -06:00
parent d7b94db0b8
commit 14631f7d78
1 changed files with 1 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
RELEASE.md
View File

@ -83,6 +83,7 @@
`tf.while_loop`, and compositions like `tf.foldl`) computed with `tf.while_loop`, and compositions like `tf.foldl`) computed with
`tf.GradientTape` inside a `tf.function`. `tf.GradientTape` inside a `tf.function`.
* Changed the default step size in `gradient_checker_v2.compute_gradients` to be exactly representable as a binary floating point numbers. This avoids poluting gradient approximations needlessly, which is some cases leads to false negatives in op gradient tests. * Changed the default step size in `gradient_checker_v2.compute_gradients` to be exactly representable as a binary floating point numbers. This avoids poluting gradient approximations needlessly, which is some cases leads to false negatives in op gradient tests.
* Extended `tf.config.experimental.enable_tensor_float_32_execution` to control Tensor-Float-32 evaluation in RNNs.
* `tf.summary`: * `tf.summary`:
* New `tf.summary.graph` allows manual write of TensorFlow graph * New `tf.summary.graph` allows manual write of TensorFlow graph