mirror of
https://github.com/zebrajr/pytorch.git
synced 2025-12-06 12:20:52 +01:00
Point to README.md#from-source instead of duplicate instructions in CONTRIBUTING.md#developing-pytorc (#91850)
**Idea:** [README.md#from-source](https://github.com/pytorch/pytorch/blob/master/README.md#from-source) should be the place that describes how I as a developer builds from source. Currently, `CONTRIBUTING.md` suggests an incomplete set of install instructions that predates those in `README.md`. This PR tries to simplify and remove a dead end from the developer onboarding funnel by pointing to [README.md#from-source](https://github.com/pytorch/pytorch/blob/master/README.md#from-source). ### Details Without touching this codebase for years I tried to build repo for local development and run unit tests. I tried to capitalise on the confusion by documenting it: 1. I go to [README.md#from-source](https://github.com/pytorch/pytorch/blob/master/README.md#from-source) 2. Since it doesn't suggest how I run unit test I follow [README.md#releases-and-contributing to ](https://github.com/pytorch/pytorch/blob/master/README.md#releases-and-contributing) to [CONTRIBUTING.md#developing-pytorch](https://github.com/pytorch/pytorch/blob/master/CONTRIBUTING.md#developing-pytorch) which is written as if it's _the_ set up dev env instruction:73e5379fab/CONTRIBUTING.md (L88-L90)But this section gives competing and incomplete install instructions that does not work for me. Ex, it doesn't mention `ninja` or `pyaml` required for `python setup.py develop`. 5. Going back to the original [README.md#from-source](https://github.com/pytorch/pytorch/blob/master/README.md#from-source) setup instructions that (mostly) worked.73e5379fab/README.md (L187)#### TODO - [x] verify that it does not break any link to other documentation [skip ci] Pull Request resolved: https://github.com/pytorch/pytorch/pull/91850 Approved by: https://github.com/ZainRizvi, https://github.com/seemethere
This commit is contained in:
Egil Martinsson
PyTorch MergeBot
parent
706aa51628
commit
b33d9e2c87
|
|
@ -4,8 +4,6 @@
|
|||
|
||||
- [Contributing to PyTorch](#contributing-to-pytorch)
|
||||
- [Developing PyTorch](#developing-pytorch)
|
||||
- [Prerequisites](#prerequisites)
|
||||
- [Instructions](#instructions)
|
||||
- [Tips and Debugging](#tips-and-debugging)
|
||||
- [Nightly Checkout & Pull](#nightly-checkout--pull)
|
||||
- [Codebase structure](#codebase-structure)
|
||||
|
|
@ -79,73 +77,37 @@ to PyTorch. For more non-technical guidance about how to contribute to
|
|||
PyTorch, see the [Contributing Guide](docs/source/community/contribution_guide.rst).
|
||||
|
||||
## Developing PyTorch
|
||||
|
||||
A full set of instructions on installing PyTorch from source is here:
|
||||
https://github.com/pytorch/pytorch#from-source
|
||||
|
||||
To develop PyTorch on your machine, here are some tips:
|
||||
|
||||
### Prerequisites
|
||||
* CMake. You can install it via `pip install cmake`
|
||||
* Python >= 3.7 (3.7.6+ recommended)
|
||||
|
||||
### Instructions
|
||||
_**Note**: If you get stuck at any step, check out the [tips and debugging](#tips-and-debugging) section below for common solutions_
|
||||
|
||||
1. Uninstall all existing PyTorch installs. You may need to run `pip
|
||||
uninstall torch` multiple times. You'll know `torch` is fully
|
||||
uninstalled when you see `WARNING: Skipping torch as it is not
|
||||
installed`. (You should only have to `pip uninstall` a few times, but
|
||||
you can always `uninstall` with `timeout` or in a loop if you're feeling
|
||||
lazy.)
|
||||
|
||||
```bash
|
||||
conda uninstall pytorch -y
|
||||
yes | pip uninstall torch
|
||||
```
|
||||
|
||||
2. Clone a copy of PyTorch from source:
|
||||
|
||||
```bash
|
||||
git clone https://github.com/pytorch/pytorch
|
||||
cd pytorch
|
||||
```
|
||||
|
||||
If you already have PyTorch from source, update it:
|
||||
|
||||
```bash
|
||||
git pull --rebase
|
||||
git submodule sync --recursive
|
||||
git submodule update --init --recursive
|
||||
```
|
||||
|
||||
If you want to have no-op incremental rebuilds (which are fast), see [Make no-op build fast](#make-no-op-build-fast) below.
|
||||
|
||||
3. Follow the instructions for [installing PyTorch from source](https://github.com/pytorch/pytorch#from-source), but instead of installing PyTorch via `python setup.py install`, use `python setup.py develop`.
|
||||
|
||||
This mode will symlink the Python files from the current local source
|
||||
tree into the Python install. This way when you modify a Python file, you
|
||||
won't need to reinstall PyTorch again and again. This is especially
|
||||
useful if you are only changing Python files.
|
||||
|
||||
For example:
|
||||
- Install local PyTorch in `develop` mode
|
||||
- modify your Python file `torch/__init__.py` (for example)
|
||||
- test functionality
|
||||
|
||||
You do not need to repeatedly install after modifying Python files (`.py`). However, you would need to reinstall
|
||||
if you modify Python interface (`.pyi`, `.pyi.in`) or non-Python files (`.cpp`, `.cc`, `.cu`, `.h`, ...).
|
||||
|
||||
In case you want to reinstall, make sure that you uninstall PyTorch
|
||||
first by running `pip uninstall torch` until you see `WARNING: Skipping
|
||||
torch as it is not installed`; next run `python setup.py clean`. After
|
||||
that, you can install in `develop` mode again.
|
||||
Follow the instructions for [installing PyTorch from source](https://github.com/pytorch/pytorch#from-source). If you get stuck when developing PyTorch on your machine, check out the [tips and debugging](#tips-and-debugging) section below for common solutions.
|
||||
|
||||
### Tips and Debugging
|
||||
|
||||
* If you want to have no-op incremental rebuilds (which are fast), see [Make no-op build fast](#make-no-op-build-fast) below.
|
||||
|
||||
* When installing with `python setup.py develop` (in contrast to `python setup.py install`) you will symlink
|
||||
the Python files from the current local source-tree into the Python install.
|
||||
This way you do not need to repeatedly install after modifying Python files (`.py`).
|
||||
However, you would need to reinstall if you modify Python interface (`.pyi`, `.pyi.in`) or
|
||||
non-Python files (`.cpp`, `.cc`, `.cu`, `.h`, ...).
|
||||
|
||||
To reinstall, first uninstall all existing PyTorch installs. You may need to run `pip
|
||||
uninstall torch` multiple times. You'll know `torch` is fully
|
||||
uninstalled when you see `WARNING: Skipping torch as it is not
|
||||
installed`. (You should only have to `pip uninstall` a few times, but
|
||||
you can always `uninstall` with `timeout` or in a loop if you're feeling
|
||||
lazy.)
|
||||
|
||||
```bash
|
||||
conda uninstall pytorch -y
|
||||
yes | pip uninstall torch
|
||||
```
|
||||
|
||||
Next run `python setup.py clean`. After that, you can install in `develop` mode again.
|
||||
|
||||
* If a commit is simple and doesn't affect any code (keep in mind that some docstrings contain code
|
||||
that is used in tests), you can add `[skip ci]` (case sensitive) somewhere in your commit message to
|
||||
[skip all build / test steps](https://github.blog/changelog/2021-02-08-github-actions-skip-pull-request-and-push-workflows-with-skip-ci/).
|
||||
Note that changing the pull request body or title on GitHub itself has no effect.
|
||||
|
||||
* If you run into errors when running `python setup.py develop`, here are some debugging steps:
|
||||
1. Run `printf '#include <stdio.h>\nint main() { printf("Hello World");}'|clang -x c -; ./a.out` to make sure
|
||||
your CMake works and can compile this simple Hello World program without errors.
|
||||
|
|
@ -166,6 +128,7 @@ that, you can install in `develop` mode again.
|
|||
```bash
|
||||
ENV_KEY1=ENV_VAL1[, ENV_KEY2=ENV_VAL2]* python setup.py develop
|
||||
```
|
||||
|
||||
* If you run into issue running `git submodule update --init --recursive`. Please try the following:
|
||||
- If you encounter an error such as
|
||||
```
|
||||
|
|
|
|||
Loading…
Reference in New Issue
Block a user