bytestring over the use of decode()

Now that it is past January 1, 2020, the code example in the README should be in Python 3

Fixing funky link. We need to show the status of the release builds out on the public CI, but those aren't actually the bits that people should use. So rather than take them to the Jenkins job and its artifacts, if they click on the badge, it will just take them to the badge icon on the public CI, which isn't terribly useful.

Removed grammatical errors.

Copying what ROCm did, for ppc64le we are providing stable
build links for TensorFlow 1.15 and 2.X releases.

- Add second Udacity course - Intro to TF Lite by TensorFlow Lite and Udacity - with a link;
- Add "from Udacity" to Intro to TF for Deep Learning's description;

Add "Intro to TensorFlow for Deep Learning" by TensorFlow and Udacity with a link

Clarify that the GPU package works with Ubuntu and Windows only - same as in this pip install guide https://www.tensorflow.org/install/pip

Imported from GitHub PR https://github.com/tensorflow/tensorflow/pull/33475

Minor change to the README, adding both TF 1.15 and 2.x releases to the list of ROCm CSB links

CC:  @sunway513
Copybara import of the project:

--
97f8bb35fbe2eace9499e2242eec54de0114a0c6 by Jeff Poznanovic <jeffrey.poznanovic@amd.com>:

Updating the README.md with ROCm CSB links

Minor change to the README, adding both TF 1.15 and 2.x releases to the list of ROCm CSB links

COPYBARA_INTEGRATE_REVIEW=https://github.com/tensorflow/tensorflow/pull/33475 from parallelo:patch-1 97f8bb35fbe2eace9499e2242eec54de0114a0c6
PiperOrigin-RevId: 275917433
Change-Id: I11ebf8c64ebc66e55316ea503ede2834a39676a4

As TensorFlow moves to 2.0, eager execution is enabled by default
so there is no need to invoke `tf.enable_eager_execution()`
explicitly.

This PR removed tf.enable_eager_execution() from README.md

Signed-off-by: Yong Tang <yong.tang.github@outlook.com>

The goal for this PR is to patch Tensorflow r2.0 release, so it would fully enable ROCm
non-xla path support.

Most of the PRs been cherry-picked in this patch have already been upstreamed in the
upstream master branch.

The following were all the related commits been cherry-picked:

Commits on Aug 20, 2019
deven-amd and sunway513
adding/updating ROCm support in the ci_build scripts
d5a0eee
deven-amd and sunway513
updating Dockerfile.rocm to pick a specific version of the rocm libra… …
e335575
deven-amd and sunway513
adding a script for testing the ROCm Community Supported Build
ae83a20

Commits on Aug 22, 2019
deven-amd and sunway513
Resolve merge conflicts for PR #31393
73ff708
deven-amd and sunway513
The following PR/commit breaks the --config=rocm build …
614bdb5
deven-amd and sunway513
updating testcases to work correctly with ROCm
1685240
jeffdaily and sunway513
improve concurrency between compute and nccl streams …
3fbb049
whchung and sunway513
[ROCm] enable roll op on ROCm.
1d5f440
whchung and sunway513
[ROCm] enable InTopK op on ROCm.
941f713
deven-amd and sunway513
updating README.md with information on ROCm Community Supported Builds
73ce64e

Commits on Aug 25, 2019
houtoms and sunway513
fixed potential rocm breaks from use_padded_io
0832b33
deven-amd and sunway513
adding no_rocm tag on unit-tests that check features that are current… …
7aed626
deven-amd and sunway513
Adding ROCm support for reduction ops
82bd216
sunway513
Fix ROCm path build error in rocm_dnn.h
5dba305

Commits on Aug 27, 2019
deven-amd
fixing test failures by skipping parts that functionality not yet sup… …
be6378c
sunway513
Merge pull request #616 from ROCmSoftwarePlatform/r2.0-rocm-upstream-… …
d98a943
sunway513
Add no_rocm tag to //tensorflow/python:stateful_random_ops_test_gpu
d05a47f

Commits on Sep 04, 2019
sunway513
Merge branch 'r2.0-rocm-upstream' of https://github.com/ROCmSoftwareP… …
b1148e4

Commits on Sep 06, 2019
deven-amd and sunway513
adding ROCm support in the build_pip_package script
b908324

PiperOrigin-RevId: 266978247

PiperOrigin-RevId: 263776602

add tensorflow in practice specialization in more info section

PiperOrigin-RevId: 256237862

PiperOrigin-RevId: 256171145

PiperOrigin-RevId: 254421795

Updated README for s390x Stable release.

Updated README for s390x Nightly artifacts.
We are working on release artifacts and will do PR for release artifacts soon.

PiperOrigin-RevId: 251269521

See https://github.com/tensorflow/community/issues/28

--

PiperOrigin-RevId: 244433092