Skip to content

Commit

Permalink
Merge branch 'develop'
Browse files Browse the repository at this point in the history
  • Loading branch information
Jopyth committed Jan 13, 2023
2 parents b125e82 + 583d59d commit d6c14d5
Show file tree
Hide file tree
Showing 137 changed files with 12,874 additions and 1,527 deletions.
67 changes: 50 additions & 17 deletions .gitlab-ci.yml
Original file line number Diff line number Diff line change
Expand Up @@ -2,41 +2,74 @@ image: python:3.7

before_script:
- python --version
- pip install -e .[dev] --extra-index-url https://download.pytorch.org/whl/cu113
- pwd
- ls -l
- python -c "import sys;print(sys.path)"
- pip install -e ".[dev]"

# jobs extending .scheduled_only only run in scheduled pipelines not on every commit
.scheduled_only:
rules:
- if: $CI_PIPELINE_SOURCE == "schedule"

# Checking codestyle

.codestyle:
stage: test
script:
- flake8 --version
- mypy --version
- black --version
- flake8
- mypy --config-file mypy.ini
- black . --check --verbose --diff --color

codestyle:
extends: .codestyle
extends:
- .codestyle

codestyle:3.8:
extends: .codestyle
image: python:3.8
codestyle:3.10:
extends:
- .codestyle
- .scheduled_only
image: python:3.10

.test:
stage: test
# Running tests

test_and_coverage:
script:
- coverage run -m pytest
- coverage report
- coverage xml
coverage: '/(?i)total.*? (100(?:\.0+)?\%|[1-9]?\d(?:\.\d+)?\%)$/'
artifacts:
reports:
coverage_report:
coverage_format: cobertura
path: coverage.xml

test:3.10:
extends:
- .scheduled_only
image: python:3.10
script:
- pytest --version
- python -m pytest .

test:
extends: .test
test:torch_backwards_compatibility:
script:
- pip install torch==1.9.0 torchvision==0.10.0
- python -m pytest .

test:3.8:
extends: .test
image: python:3.8
# Documentation

test-build-doc:
test_build_doc:
stage: test
script:
- apt-get update && apt-get install -y pandoc
- sphinx-build -b html docs/source/ docs/build/ -a

test_doc_completeness:
extends: .scheduled_only
stage: test
allow_failure: true
script:
- flake8 --version
# explicitly select Docstring errors and ignore to overwrite config in setup.cfg
- flake8 --select=D1 --ignore=E501
6 changes: 6 additions & 0 deletions .pre-commit-config.yaml
Original file line number Diff line number Diff line change
@@ -0,0 +1,6 @@
repos:
- repo: https://github.com/psf/black
rev: 22.3.0
hooks:
- id: black
language_version: python3.8
30 changes: 30 additions & 0 deletions .readthedocs.yaml
Original file line number Diff line number Diff line change
@@ -0,0 +1,30 @@
# .readthedocs.yaml
# Read the Docs configuration file
# See https://docs.readthedocs.io/en/stable/config-file/v2.html for details

# Required
version: 2

# Set the version of Python and other tools you might need
build:
os: ubuntu-20.04
tools:
python: "3.9"
# You can also specify other tool versions:
# nodejs: "16"
# rust: "1.55"
# golang: "1.17"

# Build documentation in the docs/ directory with Sphinx
sphinx:
configuration: docs/source/conf.py

# If using Sphinx, optionally build your docs in additional formats such as PDF
# formats:
# - pdf

# Optionally declare the Python requirements required to build your docs
python:
install:
- requirements: requirements.txt
- requirements: requirements-dev.txt
10 changes: 10 additions & 0 deletions AUTHORS
Original file line number Diff line number Diff line change
@@ -0,0 +1,10 @@
The development of BITorch was started by Joseph Bethge and PD Dr. Haojin Yang.

The current maintainers can be contacted at: fb10-xnor@hpi.de

The following people have contributed code to BITorch (in alphabetical order):

Christopher Aust
Joseph Bethge
Paul Mattes
Haojing Yang
38 changes: 38 additions & 0 deletions CHANGELOG.md
Original file line number Diff line number Diff line change
Expand Up @@ -4,6 +4,44 @@ All notable changes to this project will be documented in this file.
The format is based on [Keep a Changelog](http://keepachangelog.com/)
and this project adheres to [Semantic Versioning](http://semver.org/).

## [0.3.0] - 2023/01/13

### Added

- new models:
- [MeliusNet](bitorch/models/meliusnet.py)
- [BinaryDenseNet](bitorch/models/densenet.py)
- [QuickNet](bitorch/models/quicknet.py)
- simple example script for MNIST
- support for integration of bitorch's inference engine for the following layers
- QLinear
- QConv
- a quantized DLRM version, derived from [this](https://github.com/facebookresearch/dlrm) implementation
- example code for training the quantized DLRM model
- new quantization function: [Progressive Sign](bitorch/quantizations/progressive_sign.py)
- new features in PyTorch Lightning example:
- training with Knowledge Distillation
- improved logging
- callback to update Progressive Sign module
- option to integrate custom models, datasets, quantization functions
- a quantization scheduler which lets you change quantization methods during training
- a padding layer

### Changed

- requirements changed:
- code now depends on torch 1.12.x and torchvision 0.13.x
- requirements for examples are now stored at their respective folders
- optional requirements now install everything needed to run all examples
- code is now formatted with the black code formatter
- using PyTorch's implementation of RAdam
- renamed the `bitwidth` attribute of quantization functions to `bit_width`
- moved the image datasets out of the bitorch core package into the image classification example

### Fixed

- fix error from updated protobuf package

## [0.2.0] - 2022/05/19

### Added
Expand Down
65 changes: 45 additions & 20 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -9,7 +9,10 @@ Our current roadmap contains:
- Extending the model zoo with pre-trained models of state-of-the-art approaches
- Adding examples for advanced training methods with multiple stages, knowledge distillation, etc.

All changes are tracked in the [changelog](CHANGELOG.md).
All changes are tracked in the [changelog](https://github.com/hpi-xnor/bitorch/blob/main/CHANGELOG.md).

Please refer to [our wiki](https://bitorch.readthedocs.io/en/latest/) for a comprehensive introduction into
the library or use the introduction notebook in `examples/notebooks`.

## Installation

Expand All @@ -18,11 +21,11 @@ Currently, the only supported installation is pip (a conda package is planned in

### Pip

If you wish to use a *specific version* of PyTorch for compatibility with certain devices or CUDA versions,
If you wish to use a _specific version_ of PyTorch for compatibility with certain devices or CUDA versions,
we advise on installing the corresponding versions of `pytorch` and `torchvision` first (or afterwards),
please consult [pytorch's getting started guide](https://pytorch.org/get-started/locally/).

Afterwards simply run:
Otherwise, simply run:
```bash
pip install bitorch
```
Expand All @@ -32,9 +35,7 @@ Note, that you can also request a specific PyTorch version directly, e.g. for CU
pip install bitorch --extra-index-url https://download.pytorch.org/whl/cu113
```

To use advanced logging capabilities with [tensorboardX](https://github.com/lanpa/tensorboardX),
install the optional dependencies as well:

If you want to run the examples install the optional dependencies as well:
```bash
pip install "bitorch[opt]"
```
Expand All @@ -45,13 +46,8 @@ The package can also be installed locally for editing and development.
First, clone the [repository](https://github.com/hpi-xnor/bitorch), then run:

```bash
pip install -e .
```

To activate advanced logging with Tensorboard and model summary, install the optional dependencies as well:

```bash
pip install -e ".[opt]"
pip install -e . # without optional dependencies
pip install -e ".[opt]" # with optional dependencies
```

### Dali Preprocessing
Expand All @@ -61,31 +57,60 @@ e.g. with CUDA 11.x, (currently only supported for imagenet)
you need to install the `nvidia-dali-cuda110` package by running the following command:

```
pip install --extra-index-url https://developer.download.nvidia.com/compute/redist --upgrade nvidia-dali-cuda110
pip install --extra-index-url https://developer.download.nvidia.com/compute/redist --upgrade nvidia-dali-cuda110
```

### Code formatting and typing
## Development

Install the _dev_ requirements for (local) development:
Install the package and _dev_ requirements locally for development:

```bash
pip install -e ".[dev]"
```

New code should be compatible with Python 3.X versions and be compliant with PEP8. To check the codebase, please run
### Tests

The tests can be run with [pytest](https://docs.pytest.org/):

```bash
pytest
```

### Code formatting and typing

For conveniently checking whether your code suites the required style (more details below), run
```bash
./check-codestyle.sh
```

New code should be compatible with Python 3.X versions and be compliant with PEP8. To check the codebase, please run
```bash
flake8
```

The codebase has type annotations, please make sure to add type hints if required. We use `mypy` for type checking:

```bash
mypy --config-file mypy.ini
```

Finally, the tests can be run with:
For code formatting we use `black`:
```bash
black . --check --verbose --diff --color # check what changes the formatter would do
black . # apply the formatter
```

In order to automatically apply the code formatting with every commit, you can also install pre-commit
and use the pre-commit hook:
```bash
pre-commit install
```

### Documentation

We use [Google's Python Docstring Format](https://sphinxcontrib-napoleon.readthedocs.io/en/latest/example_google.html)
to document our code.

Documentation can be generated with
```bash
pytest
sphinx-build -b html docs/source/ docs/build/ -a
```
30 changes: 23 additions & 7 deletions bitorch/__init__.py
Original file line number Diff line number Diff line change
@@ -1,11 +1,18 @@
"""
BITorch is a library currently under development to simplify building quantized and binary neural networks with PyTorch.
It contains implementation of the required layers, different quantization functions and examples.
"""
import os
from argparse import ArgumentParser, Namespace
from importlib import import_module
from pathlib import Path
from typing import List

from .config import Config
from .runtime_mode import RuntimeMode, runtime_mode_type, change_mode, pause_wrapping # noqa: F401
from .layers import convert # noqa: F401

mode: RuntimeMode = RuntimeMode.DEFAULT

configs_by_name = {}

Expand Down Expand Up @@ -50,7 +57,7 @@ def config_from_name(name: str) -> Config:


def config_names() -> List:
"""getter for list of config names for argparse
"""Get the list of config names for argparse.
Returns:
List: the config names
Expand All @@ -59,20 +66,29 @@ def config_names() -> List:


def add_config_args(parser: ArgumentParser) -> None:
"""adds all config arguments
"""Adds all arguments from all registered configs.
Args:
parser (ArgumentParser): parser to add the arguments to
"""
for config in configs_by_name.values():
config.add_config_arguments(parser)
for config_ in configs_by_name.values():
config_.add_config_arguments(parser)


def apply_args_to_configuration(args: Namespace) -> None:
"""applys the cli configurations to the config objects.
"""Applies the cli configurations to the config objects.
Args:
args (Namespace): the cli configurations
"""
for config in configs_by_name.values():
config.apply_args_to_configuration(args)
for config_ in configs_by_name.values():
config_.apply_args_to_configuration(args)


def register_custom_config(custom_config: Config) -> None:
"""Register a custom (external) config in bitorch.
Args:
custom_config: the custom config which should be added to bitorch
"""
configs_by_name[custom_config.name] = custom_config
Loading

0 comments on commit d6c14d5

Please sign in to comment.