Initial commit (#2670)

Signed-off-by: Kim, Vinnam <vinnam.kim@intel.com>

.github/workflows/labeler.yaml

@@ -3,7 +3,7 @@ on:
   - pull_request_target
 
 jobs:
-  triage:
+  labeler:
     permissions:
       contents: read
       pull-requests: write

NOTICE

@@ -0,0 +1,3 @@
+This repository is under Apache License Version 2.0 licence. However the following files are borrowed from the external sources under different licenses.
+
+- MIT License: `src/otx/core/engine/**/*.py`

for_developers/dir_structure.md

@@ -0,0 +1,38 @@
+```console
+root/
+    algo/                       # Custom algo (e.g., hierarchical_cls_head)
+    cli/                        # CLI entrypoints
+    config/                     # Default YAML config files
+    core/
+        config/                 # Structured data type object for configurations
+        data/                   # Data related things
+            dataset/            # OTXDataset
+                base.py
+                detection.py
+                ...
+            entity/             # OTXDataEntity
+                base.py
+                detection.py
+                ...
+            transform_libs/     # To support transform libraries (e.g., MMCV)
+            factory.py          # Factory to instantiate data related objects
+            module.py           # OTXDataModule
+        engine/                 # PyTorchLightning engine
+            train.py
+            ...
+        model/                  # Model related things
+            entity/             # OTXModel
+                base.py
+                detection.py
+                ...
+            module/             # OTXLitModule
+                base.py
+                detection.py
+                ...
+        types/                  # Enum definitions (e.g. OTXTaskType)
+        utils/                  # Utility functions
+    recipe/                     # Recipe YAML config for each model we support
+        detection/              # (e.g., rtmdet_tiny)
+        ...
+    tools/                      # Python runnable scripts for some TBD use cases
+```

for_developers/requirements-lock.txt

@@ -0,0 +1,101 @@
+apturl==0.5.2
+attrs==23.1.0
+bandit==1.7.4
+black==23.9.1
+blinker==1.4
+Brlapi==0.8.3
+cachetools==5.3.0
+certifi==2020.6.20
+cfgv==3.4.0
+chardet==5.1.0
+click==8.0.4
+colorama==0.4.6
+command-not-found==0.3
+coverage==7.3.2
+cryptography==3.4.8
+cupshelpers==1.0
+dbus-python==1.2.18
+defer==1.0.6
+distlib==0.3.6
+distro==1.7.0
+distro-info===1.1build1
+exceptiongroup==1.1.1
+filelock==3.10.7
+gitdb==4.0.11
+GitPython==3.1.40
+httplib2==0.20.2
+identify==2.5.31
+idna==3.3
+importlib-metadata==4.6.4
+iniconfig==2.0.0
+jeepney==0.7.1
+keyring==23.5.0
+language-selector==0.1
+launchpadlib==1.10.16
+lazr.restfulclient==0.14.4
+lazr.uri==1.0.6
+louis==3.20.0
+macaroonbakery==1.3.1
+maturin==1.1.0
+more-itertools==8.10.0
+mypy==1.6.0
+mypy-extensions==1.0.0
+netifaces==0.11.0
+nodeenv==1.8.0
+oauthlib==3.2.0
+olefile==0.46
+packaging==23.0
+pathspec==0.11.2
+pbr==5.11.1
+pexpect==4.8.0
+Pillow==9.0.1
+platformdirs==2.5.1
+pluggy==1.0.0
+pre-commit==2.15.0
+protobuf==3.12.4
+ptyprocess==0.7.0
+py==1.10.0
+pycairo==1.20.1
+pycups==2.0.1
+PyGObject==3.42.1
+PyJWT==2.3.0
+pymacaroons==0.13.0
+PyNaCl==1.5.0
+pyparsing==2.4.7
+pyproject_api==1.5.1
+pyRFC3339==1.1
+pytest==6.2.5
+pytest-cov==2.11.1
+pytest-html==3.2.0
+pytest-metadata==2.0.4
+python-apt==2.3.0+ubuntu2.1
+python-dateutil==2.8.1
+python-debian===0.1.43ubuntu1
+pytz==2022.1
+pyxdg==0.27
+PyYAML==5.4.1
+reportlab==3.6.8
+requests==2.25.1
+ruff==0.0.292
+screen-resolution-extra==0.0.0
+SecretStorage==3.3.1
+six==1.16.0
+smmap==5.0.1
+ssh-import-id==5.11
+stevedore==5.1.0
+systemd-python==234
+testfixtures==7.0.0
+toml==0.10.2
+tomli==2.0.1
+tox==4.4.8
+typing_extensions==4.8.0
+ubuntu-advantage-tools==27.9
+ubuntu-drivers-common==0.0.0
+ufw==0.36.1
+unattended-upgrades==0.1
+urllib3==1.26.5
+virtualenv==20.21.0
+wadllib==1.3.6
+xdg==5
+xkit==0.0.0
+zipp==1.0.0

for_developers/setup_guide.md

@@ -0,0 +1,32 @@
+# How to setup dev env
+
+## Installation
+
+```console
+# Create venv from conda
+conda create -n otx-v2 python=3.11
+conda activate otx-v2
+
+# Install PyTorch and TorchVision
+conda install pytorch torchvision torchaudio pytorch-cuda=11.8 -c pytorch -c nvidia
+
+# Install core dependency
+pip install lightning datumaro omegaconf hydra-core
+
+# Install mmcv (mmdet)
+pip install -U openmim
+mim install mmengine "mmcv>=2.0.0" mmdet
+
+# Install this package (Sorry the installation step is not configured yet, until that please setup PYTHONPATH env var as follows)
+export PYTHONPATH=${PYTHONPATH}:${PWD}/src
+```
+
+Please see [requirements-lock.txt](requirements-lock.txt). This is what I got after the above installation steps by `pip freeze`.
+
+## Launch training with demo recipe
+
+```
+# Please check whether your PYTHONPATH is correctly setup first
+
+python src/otx/cli/train.py +recipe=detection/atss_r50_fpn base.data_dir=tests/assets/car_tree_bug model.otx_model.config.bbox_head.num_classes=3 trainer.max_epochs=50 trainer.check_val_every_n_epoch=10 trainer=gpu
+```

pyproject.toml

@@ -122,8 +122,6 @@ python_version = 3.8
 ignore_missing_imports = true
 show_error_codes = true
 
-# TODO: Need to be edited
-follow_imports = "skip"
 
 
 # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # #
@@ -191,16 +189,27 @@ ignore = [
     # pydocstyle
     # On top of the Google convention, disable `D417`, which requires
     # documentation for every function parameter.
-    "D417", # Missing argument descriptions in the docstring
+    "D417",     # Missing argument descriptions in the docstring
+
+    "D107",     # Missing docstring in `__init__`
+    "D105",     # Missing docstring in magic method
 
     # flake8-annotations
     "ANN101",   # Missing-type-self
     "ANN002",   # Missing type annotation for *args
     "ANN003",   # Missing type annotation for **kwargs
+
+    "ANN102",   # Missing type annotation for `cls` in classmethod
+    "ANN204",   # Missing return type annotation for special method `__init__`
+
+    "ARG002",   # Unused method argument
+
     # flake8-type-checking
     "TCH001",   # typing-only-first-party-import, Sometimes this causes an incorrect error.
     # flake8-fixme
-    "FIX002"    # line-contains-todo
+    "FIX002",   # line-contains-todo
+
+    "UP007"     # Use `X | Y` for type annotations
 ]
 
 # Allow autofix for all enabled rules (when `--fix`) is provided.
@@ -232,6 +241,9 @@ exclude = [
     "dist",
     "node_modules",
     "venv",
+    "tests/assets",
+
+    "src/otx/core/engine/utils/*"   # It will be clean up later
 ]
 
 # Same as Black.
@@ -265,6 +277,9 @@ max-returns = 10
     "ANN001",   # Skip annotation type hint in test codes
     "D",     # Test skips missing docstring argument with magic (fixture) methods.
 ]
+"src/otx/**/*.py" = [
+    "ERA001",
+]
 
 [tool.ruff.pydocstyle]
 convention = "google"

src/otx/algo/__init__.py

@@ -0,0 +1,4 @@
+# Copyright (C) 2023 Intel Corporation
+# SPDX-License-Identifier: Apache-2.0
+#
+"""Module for OTX custom algorithms, e.g., model, losses, hook, etc..."""

src/otx/algo/model/__init__.py

@@ -0,0 +1,4 @@
+# Copyright (C) 2023 Intel Corporation
+# SPDX-License-Identifier: Apache-2.0
+#
+"""Module for OTX custom model."""

src/otx/algo/model/hierarchical_cls_head.py

@@ -0,0 +1,4 @@
+# Copyright (C) 2023 Intel Corporation
+# SPDX-License-Identifier: Apache-2.0
+#
+"""Hierarchical classification head PyTorch module."""

src/otx/cli/__init__.py

@@ -0,0 +1,4 @@
+# Copyright (C) 2023 Intel Corporation
+# SPDX-License-Identifier: Apache-2.0
+#
+"""CLI entrypoints."""

src/otx/cli/train.py

@@ -0,0 +1,41 @@
+# Copyright (C) 2023 Intel Corporation
+# SPDX-License-Identifier: Apache-2.0
+#
+"""CLI entrypoint for training."""
+# ruff: noqa
+
+import hydra
+from omegaconf import DictConfig
+
+from otx.core.config import register_configs
+
+register_configs()
+
+
+@hydra.main(version_base="1.3", config_path="../config", config_name="train.yaml")
+def main(cfg: DictConfig) -> None:
+    """Main entry point for training.
+
+    :param cfg: DictConfig configuration composed by Hydra.
+    :return: Optional[float] with optimized metric value.
+    """
+    from otx.core.engine.train import train
+
+    # apply extra utilities
+    # (e.g. ask for tags if none are provided in cfg, print cfg tree, etc.)
+    # utils.extras(cfg)
+
+    # train the model
+    metric_dict, _ = train(cfg)
+
+    # # safely retrieve metric value for hydra-based hyperparameter optimization
+    # metric_value = utils.get_metric_value(
+    #     metric_dict=metric_dict, metric_name=cfg.get("optimized_metric")
+    # )
+
+    # # return optimized metric
+    # return metric_value
+
+
+if __name__ == "__main__":
+    main()

src/otx/config/__init__.py

@@ -0,0 +1,4 @@
+"""OTX default YAML configuration file collection.
+
+this file is needed here to include configs when building project as a package
+"""

src/otx/config/base/default.yaml

@@ -0,0 +1,22 @@
+# OTX task enum
+task: ???
+
+# # simply provide checkpoint path to resume training
+# ckpt_path: null
+
+# # seed for random number generators in pytorch, numpy and python.random
+# seed: null
+
+# path to working directory
+work_dir: ${hydra:runtime.cwd}
+
+# path to data directory
+data_dir: ${base.work_dir}/data/
+
+# path to logging directory
+log_dir: ${base.work_dir}/logs/
+
+# path to output directory, created dynamically by hydra
+# path generation pattern is specified in `configs/hydra/default.yaml`
+# use it to store all files generated during the run, like ckpts and metrics
+output_dir: ${hydra:runtime.output_dir}

src/otx/config/base/detection.yaml

@@ -0,0 +1,4 @@
+defaults:
+  - default
+
+task: DETECTION

src/otx/config/callbacks/default.yaml

@@ -0,0 +1,22 @@
+defaults:
+  - model_checkpoint
+  - early_stopping
+  - model_summary
+  - rich_progress_bar
+  - _self_
+
+model_checkpoint:
+  dirpath: ${paths.output_dir}/checkpoints
+  filename: "epoch_{epoch:03d}"
+  monitor: "val/acc"
+  mode: "max"
+  save_last: True
+  auto_insert_metric_name: False
+
+early_stopping:
+  monitor: "val/acc"
+  patience: 100
+  mode: "max"
+
+model_summary:
+  max_depth: -1

src/otx/config/callbacks/early_stopping.yaml

@@ -0,0 +1,15 @@
+# https://lightning.ai/docs/pytorch/stable/api/lightning.pytorch.callbacks.EarlyStopping.html
+
+early_stopping:
+  _target_: lightning.pytorch.callbacks.EarlyStopping
+  monitor: ??? # quantity to be monitored, must be specified !!!
+  min_delta: 0. # minimum change in the monitored quantity to qualify as an improvement
+  patience: 3 # number of checks with no improvement after which training will be stopped
+  verbose: False # verbosity mode
+  mode: "min" # "max" means higher metric value is better, can be also "min"
+  strict: True # whether to crash the training if monitor is not found in the validation metrics
+  check_finite: True # when set True, stops training when the monitor becomes NaN or infinite
+  stopping_threshold: null # stop training immediately once the monitored quantity reaches this threshold
+  divergence_threshold: null # stop training as soon as the monitored quantity becomes worse than this threshold
+  check_on_train_epoch_end: null # whether to run early stopping at the end of the training epoch
+  # log_rank_zero_only: False  # this keyword argument isn't available in stable version

src/otx/config/callbacks/model_checkpoint.yaml

@@ -0,0 +1,17 @@
+# https://lightning.ai/docs/pytorch/stable/api/lightning.pytorch.callbacks.ModelCheckpoint.html
+
+model_checkpoint:
+  _target_: lightning.pytorch.callbacks.ModelCheckpoint
+  dirpath: null # directory to save the model file
+  filename: null # checkpoint filename
+  monitor: null # name of the logged metric which determines when model is improving
+  verbose: False # verbosity mode
+  save_last: null # additionally always save an exact copy of the last checkpoint to a file last.ckpt
+  save_top_k: 1 # save k best models (determined by above metric)
+  mode: "min" # "max" means higher metric value is better, can be also "min"
+  auto_insert_metric_name: True # when True, the checkpoints filenames will contain the metric name
+  save_weights_only: False # if True, then only the model’s weights will be saved
+  every_n_train_steps: null # number of training steps between checkpoints
+  train_time_interval: null # checkpoints are monitored at the specified time interval
+  every_n_epochs: null # number of epochs between checkpoints
+  save_on_train_epoch_end: null # whether to run checkpointing at the end of the training epoch or the end of validation