Docs welcome page and Development / Inference page overhaul #1793

JRMeyer · 2021-03-14T16:08:50Z

This is a purely document-related PR

Streamlined the welcome page for readthedocs (i.e. index.rst)
modified and re-ordered DEPLOYMENT.rst

reuben

Nice! r+ with comments addressed.

reuben · 2021-03-14T16:26:27Z

doc/DEPLOYMENT.rst

+* :ref:`The Node.JS package + language binding <nodejs-usage>`
+* :github:`The .NET client + language binding <native_client/dotnet/README.rst>`
+
+Download trained 🐸STT models


I've been using the policy of expanding 🐸 -> Coqui in headings in the docs.

reuben · 2021-03-14T16:27:00Z

doc/SUPPORTED_PLATFORMS.rst


-Here we maintain the list of supported platforms for running inference. Note that for now we only have working packages for Python on Linux, without GPU support. We're working to get the rest of our supported languages and architectures up and running.
+Here we maintain the list of supported platforms for running inference.


We should be consistent on deployment/inference throughout the docs.

reuben · 2021-03-14T16:28:55Z

doc/index.rst


-   BUILDING_DotNet
+The fastest way to deploy a pre-trained 🐸STT model is with `pip`: 


We should retain the note about Linux only support for now here to avoid people running into missing packages as after following the first explicit step-by-step instructions in our docs.

reuben · 2021-03-14T16:29:30Z

doc/index.rst


-   Scorer
+   # Create a virtual environment
+   $ python -m venv venv-stt


Suggested change

$ python -m venv venv-stt

$ python3 -m venv venv-stt

We don't support Python 2.

python on any modern system is python3. python2 is long past it's sunset date [1].

So, I think it's better to say python than python3. I specify python>=3.5 in DEPLOYMENT.rst [2] , but you're right, it should be explicit in index.rst

Bare python executable is still Python 2 on lots of running systems. That's also up to the distro and not Python.org's support timeline. So it's better to be safe here and explicitly create the venv with the Python 3 executable.

doc/index.rst

reuben · 2021-03-14T16:32:13Z

doc/DEPLOYMENT.rst

+
+.. code-block::
+
+   (coqui-stt-venv)$ python -m pip install stt


Suggested change

(coqui-stt-venv)$ python -m pip install stt

(coqui-stt-venv)$ pip install -U pip && pip install stt

reuben · 2021-03-14T16:33:35Z

doc/DEPLOYMENT.rst

+
+After installation has finished, you can call ``stt`` from the command-line.
+
+The following command assumes you `downloaded the pre-trained models <#getting-the-pre-trained-model>`_.


The correct way to reference sections in RST is by giving them a name, so that this link doesn't break if you change the section heading. See for example the py-usage section name above or py-api-example just below.

reuben · 2021-03-14T16:35:09Z

doc/DEPLOYMENT.rst

+
+Alternatively you may manually download the ``native_client.tar.xz`` from the `releases page <https://github.com/coqui-ai/STT/releases>`_.
+
+Assuming you have `downloaded the pre-trained models <#getting-the-pre-trained-model>`_, you can use the client as such: 


Use a proper section name for linking.

reuben · 2021-03-14T16:35:57Z

doc/DEPLOYMENT.rst

+
+See :ref:`the Python client <py-api-example>` for an example of how to use the package programatically.
+
+*GPUs will soon be supported:* If you have a supported NVIDIA GPU on Linux, you can install the GPU specific package as follows:


If we're adding a note about missing GPU support here then should we also add a note about missing JS support in the JS docs? And missing GPU support in the GPU docs for JS?

reuben · 2021-03-14T16:36:39Z

doc/DEPLOYMENT.rst

+Installing bindings from source
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+If pre-built binaries aren't available for your system, you'll need to install them from scratch. Follow the :github:`native client build and installation instructions <native_client/README.rst>`.


This link is broken, should be:

Suggested change

If pre-built binaries aren't available for your system, you'll need to install them from scratch. Follow the :github:`native client build and installation instructions <native_client/README.rst>`.

If pre-built binaries aren't available for your system, you'll need to install them from scratch. Follow the :github:`native client build and installation instructions <build-native-client>`.

Docs welcome page and Development / Inference page overhaul

4d8e579

reuben approved these changes Mar 14, 2021

View reviewed changes

JRMeyer and others added 2 commits March 16, 2021 18:50

reubens reviews

36a8da9

Fix broken refs and other small adjustments

a25356a

reuben merged commit 629706b into coqui-ai:main Mar 17, 2021

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Docs welcome page and Development / Inference page overhaul #1793

Docs welcome page and Development / Inference page overhaul #1793


		Here we maintain the list of supported platforms for running inference. Note that for now we only have working packages for Python on Linux, without GPU support. We're working to get the rest of our supported languages and architectures up and running.
		Here we maintain the list of supported platforms for running inference.


		BUILDING_DotNet
		The fastest way to deploy a pre-trained 🐸STT model is with `pip`:

	(coqui-stt-venv)$ python -m pip install stt
	(coqui-stt-venv)$ pip install -U pip && pip install stt


		After installation has finished, you can call ``stt`` from the command-line.

		The following command assumes you `downloaded the pre-trained models <#getting-the-pre-trained-model>`_.


		Alternatively you may manually download the ``native_client.tar.xz`` from the `releases page <https://github.com/coqui-ai/STT/releases>`_.

		Assuming you have `downloaded the pre-trained models <#getting-the-pre-trained-model>`_, you can use the client as such:


		See :ref:`the Python client <py-api-example>` for an example of how to use the package programatically.

		GPUs will soon be supported: If you have a supported NVIDIA GPU on Linux, you can install the GPU specific package as follows:

	If pre-built binaries aren't available for your system, you'll need to install them from scratch. Follow the :github:`native client build and installation instructions <native_client/README.rst>`.
	If pre-built binaries aren't available for your system, you'll need to install them from scratch. Follow the :github:`native client build and installation instructions <build-native-client>`.

Docs welcome page and Development / Inference page overhaul #1793

Docs welcome page and Development / Inference page overhaul #1793

Conversation

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment