LMFuser 0.0.1__tar.gz

lmfuser-0.0.1/.github/workflows/python-publish.yml

@@ -0,0 +1,70 @@
+# This workflow will upload a Python Package to PyPI when a release is created
+# For more information see: https://docs.github.com/en/actions/automating-builds-and-tests/building-and-testing-python#publishing-to-package-registries
+
+# This workflow uses actions that are not certified by GitHub.
+# They are provided by a third-party and are governed by
+# separate terms of service, privacy policy, and support
+# documentation.
+
+name: Upload Python Package
+
+on:
+  release:
+    types: [published]
+
+permissions:
+  contents: read
+
+jobs:
+  release-build:
+    runs-on: ubuntu-latest
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.11"
+
+      - name: Build release distributions
+        run: |
+          # NOTE: put your own distribution build steps here.
+          python -m pip install build
+          python -m build
+
+      - name: Upload distributions
+        uses: actions/upload-artifact@v4
+        with:
+          name: release-dists
+          path: dist/
+
+  pypi-publish:
+    runs-on: ubuntu-latest
+    needs:
+      - release-build
+    permissions:
+      # IMPORTANT: this permission is mandatory for trusted publishing
+      id-token: write
+
+    # Dedicated environments with protections for publishing are strongly recommended.
+    # For more information, see: https://docs.github.com/en/actions/deployment/targeting-different-environments/using-environments-for-deployment#deployment-protection-rules
+    environment:
+      name: pypi
+      # OPTIONAL: uncomment and update to include your PyPI project URL in the deployment status:
+      # url: https://pypi.org/p/YOURPROJECT
+      #
+      # ALTERNATIVE: if your GitHub Release name is the PyPI project version string
+      # ALTERNATIVE: exactly, uncomment the following line instead:
+      # url: https://pypi.org/project/YOURPROJECT/${{ github.event.release.name }}
+
+    steps:
+      - name: Retrieve release distributions
+        uses: actions/download-artifact@v4
+        with:
+          name: release-dists
+          path: dist/
+
+      - name: Publish release distributions to PyPI
+        uses: pypa/gh-action-pypi-publish@release/v1
+        with:
+          packages-dir: dist/

lmfuser-0.0.1/.gitignore

@@ -0,0 +1,207 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[codz]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py.cover
+.hypothesis/
+.pytest_cache/
+cover/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+.pybuilder/
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+#   For a library or package, you might want to ignore these files since the code is
+#   intended to run in multiple environments; otherwise, check them in:
+# .python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# UV
+#   Similar to Pipfile.lock, it is generally recommended to include uv.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#uv.lock
+
+# poetry
+#   Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#   https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
+#poetry.lock
+#poetry.toml
+
+# pdm
+#   Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
+#   pdm recommends including project-wide configuration in pdm.toml, but excluding .pdm-python.
+#   https://pdm-project.org/en/latest/usage/project/#working-with-version-control
+#pdm.lock
+#pdm.toml
+.pdm-python
+.pdm-build/
+
+# pixi
+#   Similar to Pipfile.lock, it is generally recommended to include pixi.lock in version control.
+#pixi.lock
+#   Pixi creates a virtual environment in the .pixi directory, just like venv module creates one
+#   in the .venv directory. It is recommended not to include this directory in version control.
+.pixi
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.envrc
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# pytype static type analyzer
+.pytype/
+
+# Cython debug symbols
+cython_debug/
+
+# PyCharm
+#  JetBrains specific template is maintained in a separate JetBrains.gitignore that can
+#  be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
+#  and can be added to the global gitignore or merged into this file.  For a more nuclear
+#  option (not recommended) you can uncomment the following to ignore the entire idea folder.
+#.idea/
+
+# Abstra
+# Abstra is an AI-powered process automation framework.
+# Ignore directories containing user credentials, local state, and settings.
+# Learn more at https://abstra.io/docs
+.abstra/
+
+# Visual Studio Code
+#  Visual Studio Code specific template is maintained in a separate VisualStudioCode.gitignore 
+#  that can be found at https://github.com/github/gitignore/blob/main/Global/VisualStudioCode.gitignore
+#  and can be added to the global gitignore or merged into this file. However, if you prefer, 
+#  you could uncomment the following to ignore the entire vscode folder
+# .vscode/
+
+# Ruff stuff:
+.ruff_cache/
+
+# PyPI configuration file
+.pypirc
+
+# Cursor
+#  Cursor is an AI-powered code editor. `.cursorignore` specifies files/directories to
+#  exclude from AI features like autocomplete and code analysis. Recommended for sensitive data
+#  refer to https://docs.cursor.com/context/ignore-files
+.cursorignore
+.cursorindexingignore
+
+# Marimo
+marimo/_static/
+marimo/_lsp/
+__marimo__/

lmfuser-0.0.1/.vscode/settings.json

@@ -0,0 +1,7 @@
+{
+    "python.analysis.typeCheckingMode": "standard",
+    "editor.rulers": [
+        80,
+        120
+    ]
+}

lmfuser-0.0.1/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2025 Yintao Tai
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

lmfuser-0.0.1/PKG-INFO

@@ -0,0 +1,30 @@
+Metadata-Version: 2.4
+Name: LMFuser
+Version: 0.0.1
+Summary: The LMFuser training framework.
+Project-URL: Homepage, https://github.com/TYTTYTTYT/LMFuser
+Project-URL: Documentation, https://github.com/TYTTYTTYT/LMFuser
+Project-URL: Repository, https://github.com/TYTTYTTYT/LMFuser
+Project-URL: Bug Tracker, https://github.com/TYTTYTTYT/LMFuser
+Project-URL: Changelog, https://github.com/TYTTYTTYT/LMFuser
+Author-email: Yintao Tai <tai.yintao@gmail.com>
+Maintainer-email: Yintao Tai <tai.yintao@gmail.com>
+License-File: LICENSE
+Keywords: machine learning,pytorch
+Classifier: Intended Audience :: Information Technology
+Classifier: Intended Audience :: Science/Research
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Topic :: Utilities
+Requires-Python: >=3.11
+Requires-Dist: hyperargs>=0.1.3
+Requires-Dist: lmfuser-data
+Requires-Dist: numpy
+Requires-Dist: pandas
+Requires-Dist: pyarrow
+Requires-Dist: torch
+Requires-Dist: tqdm
+Requires-Dist: wandb
+Description-Content-Type: text/markdown
+
+# LMFuser
+A robust, multi-task PyTorch framework for training versatile Language Models at scale.

lmfuser-0.0.1/README.md

@@ -0,0 +1,2 @@
+# LMFuser
+A robust, multi-task PyTorch framework for training versatile Language Models at scale.

lmfuser-0.0.1/pyproject.toml

@@ -0,0 +1,43 @@
+[build-system]
+requires = ["hatchling >= 1.26"]
+build-backend = "hatchling.build"
+
+[project]
+name = "LMFuser"
+version = "0.0.1"
+requires-python = ">= 3.11"
+description = "The LMFuser training framework."
+readme = "README.md"
+keywords = ["machine learning", "pytorch"]
+classifiers = [
+  "Topic :: Utilities",
+  "Programming Language :: Python :: 3 :: Only",
+  "Intended Audience :: Science/Research",
+  "Intended Audience :: Information Technology"
+]
+dependencies = [
+  "numpy",
+  "pandas",
+  "pyarrow",
+  "torch",
+  "LMFuser-data",
+  "tqdm",
+  "wandb",
+  "HyperArgs>=0.1.3",
+]
+authors = [
+  {name = "Yintao Tai", email = "tai.yintao@gmail.com"},
+]
+maintainers = [
+  {name = "Yintao Tai", email = "tai.yintao@gmail.com"},
+]
+
+[project.urls]
+Homepage = "https://github.com/TYTTYTTYT/LMFuser"
+Documentation = "https://github.com/TYTTYTTYT/LMFuser"
+Repository = "https://github.com/TYTTYTTYT/LMFuser"
+"Bug Tracker" = "https://github.com/TYTTYTTYT/LMFuser"
+Changelog = "https://github.com/TYTTYTTYT/LMFuser"
+
+[tool.hatch.build.targets.wheel]
+packages = ["src/lmfuser"]

lmfuser-0.0.1/src/lmfuser/model_loader.py

@@ -0,0 +1,33 @@
+from os import PathLike
+
+from torch import nn
+from lmfuser_data.interfaces import SubclassTracer
+from hyperargs import Conf, StrArg, OptionArg
+
+
+class ModelLoader(SubclassTracer):
+    def __init__(self, model_path: str | PathLike):
+        self.model_path = model_path
+
+    def load_model(self) -> nn.Module:
+        raise NotImplementedError(f'Not implemented load_model for {self.__class__.__name__}')
+
+    @classmethod
+    def save_model(cls, model: nn.Module, directory: str | PathLike) -> None:
+        raise NotImplementedError(f'Not implemented save_model for {model.__class__.__name__}')
+
+def find_model_loader_names() -> list[str]:
+    return list(ModelLoader.all_subclass_names()) + ['ModelLoader']
+
+
+class ModelLoaderConf(Conf):
+    model_path = StrArg('please set model path here!')
+    model_type = OptionArg(default='ModelLoader', option_fn=find_model_loader_names)
+
+    def get_model_loader(self) -> ModelLoader:
+        name = self.model_type.value()
+        path = self.model_path.value()
+        assert name is not None and path is not None
+        if name == 'ModelLoader':
+            return ModelLoader(path)
+        return ModelLoader.all_subclass_map()[name](path)

lmfuser-0.0.1/src/lmfuser/optimizers.py

@@ -0,0 +1,162 @@
+from typing import Iterable
+
+import torch
+from torch.nn.parameter import Parameter as Parameter
+from torch.optim import (
+    Optimizer,
+    Adam,
+    AdamW,
+    SGD,
+    Adadelta,
+    Adagrad
+)
+from hyperargs import Conf, OptionArg, FloatArg, BoolArg, add_dependency, monitor_on
+
+
+class OptimizerConfigBase(Conf):
+
+    def init_optimzier(
+        self,
+        params: Iterable[Parameter],
+    ) -> Optimizer:
+        raise NotImplementedError(
+            'Please Implement this method in child classes!'
+        )
+
+
+class AdamWConfig(OptimizerConfigBase):
+
+    lr = FloatArg(5e-5, min_value=0.0)
+    beta1 = FloatArg(0.9, min_value=0.0, max_value=1.0)
+    beta2 = FloatArg(0.999, min_value=0.0, max_value=1.0)
+    eps = FloatArg(1e-8, min_value=0.0, max_value=1.0)
+    weight_decay = FloatArg(0.01, min_value=0.0, max_value=1.0)
+    amsgrad = BoolArg(False)
+
+    def init_optimzier(
+        self, 
+        params: Iterable[Parameter],
+    ) -> Optimizer:
+        return AdamW(
+            params=params,
+            lr=self.lr.value(), # type: ignore
+            betas=(self.beta1.value(), self.beta2.value()), # type: ignore
+            eps=self.eps.value(), # type: ignore
+            weight_decay=self.weight_decay.value(), # type: ignore
+            amsgrad=self.amsgrad.value() # type: ignore
+        )
+
+
+class AdamConfig(OptimizerConfigBase):
+
+    lr = FloatArg(5e-5, min_value=0.0)
+    beta1 = FloatArg(0.9, min_value=0.0, max_value=1.0)
+    beta2 = FloatArg(0.999, min_value=0.0, max_value=1.0)
+    eps = FloatArg(1e-8, min_value=0.0, max_value=1.0)
+    weight_decay = FloatArg(0.01, min_value=0.0, max_value=1.0)
+    amsgrad = BoolArg(False)
+
+    def init_optimzier(
+        self, 
+        params: Iterable[Parameter]
+    ) -> Optimizer:
+        return Adam(
+            params=params,
+            lr=self.lr.value(), # type: ignore
+            betas=(self.beta1.value(), self.beta2.value()), # type: ignore
+            eps=self.eps.value(), # type: ignore
+            weight_decay=self.weight_decay.value(), # type: ignore
+            amsgrad=self.amsgrad.value() # type: ignore
+        )
+
+
+class SGDConfig(OptimizerConfigBase):
+
+    lr = FloatArg(5e-5, min_value=0.0)
+    momentum = FloatArg(0.90, min_value=0.0, max_value=1.0)
+    dampening = FloatArg(0.0, min_value=0.0, max_value=1.0)
+    weight_decay = FloatArg(0.0, min_value=0.0, max_value=1.0)
+    nesterov = BoolArg(False)
+
+    def init_optimzier(
+        self, 
+        params: Iterable[Parameter],
+    ) -> Optimizer:
+        return SGD(
+            params=params,
+            lr=self.lr.value(), # type: ignore
+            momentum=self.momentum.value(), # type: ignore
+            weight_decay=self.weight_decay.value(), # type: ignore
+            nesterov=self.nesterov.value() # type: ignore
+        )
+
+
+class AdadeltaConfig(OptimizerConfigBase):
+
+    lr = FloatArg(1.0, min_value=0.0)
+    rho = FloatArg(0.9, min_value=0.0, max_value=1.0)
+    eps = FloatArg(1e-6, min_value=0.0, max_value=1.0)
+    weight_decay = FloatArg(0.0, min_value=0.0, max_value=1.0)
+
+    def init_optimzier(
+        self, 
+        params: Iterable[Parameter]
+    ) -> Optimizer:
+        return Adadelta(
+            params=params,
+            lr=self.lr.value(), # type: ignore
+            rho=self.rho.value(), # type: ignore
+            eps=self.eps.value(), # type: ignore
+            weight_decay=self.weight_decay.value() # type: ignore
+        )
+
+
+class AdagradConfig(OptimizerConfigBase):
+
+    lr = FloatArg(0.01, min_value=0.0)
+    lr_decay = FloatArg(0.0, min_value=0.0, max_value=1.0)
+    weight_decay = FloatArg(0.0, min_value=0.0, max_value=1.0)
+    initial_accumulator_value = FloatArg(0.0, min_value=0.0, max_value=1.0)
+    eps = FloatArg(1e-10, min_value=0.0, max_value=1.0)
+
+    def init_optimzier(
+        self, 
+        params: Iterable[Parameter]
+    ) -> Optimizer:
+        return Adagrad(
+            params=params,
+            lr=self.lr.value(), # type: ignore
+            lr_decay=self.lr_decay.value(), # type: ignore
+            weight_decay=self.weight_decay.value(), # type: ignore
+            initial_accumulator_value=self.initial_accumulator_value.value(), # type: ignore
+            eps=self.eps.value() # type: ignore
+        )
+
+
+@add_dependency('optimizer_type', 'optimizer')
+class OptimizerConfig(Conf):
+    optimizer_type = OptionArg('AdamW', options=['AdamW', 'Adam', 'SGD', 'Adadelta', 'Adagrad'])
+    optimizer: OptimizerConfigBase = AdamWConfig()
+
+    @monitor_on('optimizer_type')
+    def set_optimizer(self) -> None:
+        if self.optimizer_type.value() == 'AdamW' and not isinstance(self.optimizer, AdamWConfig):
+            self.optimizer = AdamWConfig()
+        elif self.optimizer_type.value() == 'Adam' and not isinstance(self.optimizer, AdamConfig):
+            self.optimizer = AdamConfig()
+        elif self.optimizer_type.value() == 'SGD' and not isinstance(self.optimizer, SGDConfig):
+            self.optimizer = SGDConfig()
+        elif self.optimizer_type.value() == 'Adadelta' and not isinstance(self.optimizer, AdadeltaConfig):
+            self.optimizer = AdadeltaConfig()
+        elif self.optimizer_type.value() == 'Adagrad' and not isinstance(self.optimizer, AdagradConfig):
+            self.optimizer = AdagradConfig()
+
+    def init_optimzier(
+        self,
+        params: Iterable[Parameter]
+    ) -> Optimizer:
+        return self.optimizer.init_optimzier(params)
+
+if __name__ == '__main__':
+    conf = OptimizerConfig.parse_command_line()
+    print(conf)

lmfuser-0.0.1/src/lmfuser/runners/__init__.py

@@ -0,0 +1 @@
+from .ddp_runner import DDPRunner, DDPRunnerConfig