Merge branch 'pip_package' into 'main'

Add pip package for megatron.core

See merge request ADLR/megatron-lm!598

megatron/core/README.md

@@ -0,0 +1 @@
+Megatron Core is a library for efficient and scalable training of transformer based models.

megatron/core/package_info.py

@@ -0,0 +1,23 @@
+# Copyright (c) 2023, NVIDIA CORPORATION.  All rights reserved.
+
+
+MAJOR = 0
+MINOR = 1
+PATCH = 0
+PRE_RELEASE = ''
+
+# Use the following formatting: (major, minor, patch, pre-release)
+VERSION = (MAJOR, MINOR, PATCH, PRE_RELEASE)
+
+__shortversion__ = '.'.join(map(str, VERSION[:3]))
+__version__ = '.'.join(map(str, VERSION[:3])) + ''.join(VERSION[3:])
+
+__package_name__ = 'megatron_core'
+__contact_names__ = 'NVIDIA'
+__contact_emails__ = 'nemo-toolkit@nvidia.com' # use NeMo Email
+__homepage__ = 'https://docs.nvidia.com/deeplearning/nemo/user-guide/docs/en/stable/' # use NeMo homepage
+__repository_url__ = 'https://github.com/NVIDIA/Megatron-LM/megatron/core'
+__download_url__ = 'https://github.com/NVIDIA/Megatron-LM/releases'
+__description__ = 'Megatron Core - a library for efficient and scalable training of transformer based models'
+__license__ = 'BSD-3'
+__keywords__ = 'deep learning, machine learning, gpu, NLP, NLU, language, transformer, nvidia, pytorch, torch'

megatron/core/requirements.txt

@@ -0,0 +1 @@
+torch

megatron/core/utils.py

@@ -1,3 +1,5 @@
+# Copyright (c) 2023, NVIDIA CORPORATION. All rights reserved.
+
 """Utility functions used throughout Megatron core"""
 from functools import reduce
 import operator

setup.py

@@ -1,10 +1,111 @@
 from setuptools import setup, find_packages
 
-setup(
-    name="megatron.core",
-    version="0.1",
-    description="Core components of Megatron.",
-    packages=find_packages(
-        include=("megatron.core")
-    )
-)
+"""Setup for pip package."""
+
+import importlib.util
+import os
+import setuptools
+
+spec = importlib.util.spec_from_file_location('package_info', 'megatron/core/package_info.py')
+package_info = importlib.util.module_from_spec(spec)
+spec.loader.exec_module(package_info)
+
+
+__contact_emails__ = package_info.__contact_emails__
+__contact_names__ = package_info.__contact_names__
+__description__ = package_info.__description__
+__download_url__ = package_info.__download_url__
+__homepage__ = package_info.__homepage__
+__keywords__ = package_info.__keywords__
+__license__ = package_info.__license__
+__package_name__ = package_info.__package_name__
+__repository_url__ = package_info.__repository_url__
+__version__ = package_info.__version__
+
+
+if os.path.exists('megatron/core/README.md'):
+    with open("megatron/core/README.md", "r", encoding='utf-8') as fh:
+        long_description = fh.read()
+    long_description_content_type = "text/markdown"
+
+else:
+    long_description = 'See ' + __homepage__
+    long_description_content_type = "text/plain"
+
+
+###############################################################################
+#                             Dependency Loading                              #
+# %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% #
+
+def req_file(filename, folder="megatron/core"):
+    with open(os.path.join(folder, filename), encoding='utf-8') as f:
+        content = f.readlines()
+    # you may also want to remove whitespace characters
+    # Example: `\n` at the end of each line
+    return [x.strip() for x in content]
+
+install_requires = req_file("requirements.txt")
+
+###############################################################################
+
+setuptools.setup(
+    name=__package_name__,
+    # Versions should comply with PEP440.  For a discussion on single-sourcing
+    # the version across setup.py and the project code, see
+    # https://packaging.python.org/en/latest/single_source_version.html
+    version=__version__,
+    description=__description__,
+    long_description=long_description,
+    long_description_content_type=long_description_content_type,
+    # The project's main homepage.
+    url=__repository_url__,
+    download_url=__download_url__,
+    # Author details
+    author=__contact_names__,
+    author_email=__contact_emails__,
+    # maintainer Details
+    maintainer=__contact_names__,
+    maintainer_email=__contact_emails__,
+    # The licence under which the project is released
+    license=__license__,
+    classifiers=[
+        # How mature is this project? Common values are
+        #  1 - Planning
+        #  2 - Pre-Alpha
+        #  3 - Alpha
+        #  4 - Beta
+        #  5 - Production/Stable
+        #  6 - Mature
+        #  7 - Inactive
+        'Development Status :: 5 - Production/Stable',
+        # Indicate who your project is intended for
+        'Intended Audience :: Developers',
+        'Intended Audience :: Science/Research',
+        'Intended Audience :: Information Technology',
+        # Indicate what your project relates to
+        'Topic :: Scientific/Engineering',
+        'Topic :: Scientific/Engineering :: Mathematics',
+        'Topic :: Scientific/Engineering :: Image Recognition',
+        'Topic :: Scientific/Engineering :: Artificial Intelligence',
+        'Topic :: Software Development :: Libraries',
+        'Topic :: Software Development :: Libraries :: Python Modules',
+        'Topic :: Utilities',
+        # Pick your license as you wish (should match "license" above)
+        'License :: OSI Approved :: BSD License',
+        # Supported python versions
+        'Programming Language :: Python :: 3',
+        'Programming Language :: Python :: 3.8',
+        'Programming Language :: Python :: 3.9',
+        # Additional Setting
+        'Environment :: Console',
+        'Natural Language :: English',
+        'Operating System :: OS Independent',
+    ],
+    packages=['megatron.core', 'megatron.core.pipeline_parallel', 'megatron.core.tensor_parallel'], 
+    install_requires=install_requires,
+
+    # Add in any packaged data.
+    include_package_data=True,
+    # PyPI package information.
+    keywords=__keywords__,
+)