Merge pull request #10 from lombokai/dev

Merge to main from dev after developing minimun standard of machine learning project

Author: nunenuh

test/__init__.py renamed to .github/workflows/readme.md

@@ -1,12 +1,13 @@
 # ignore dataset
 raw_data/*
-src/data/train/*
-src/data/val/*
+data/train/*
+data/val/*
 
 # Byte-compiled / optimized / DLL files
 __pycache__/
 *.py[cod]
 *$py.class
+.vscode
 
 # C extensions
 *.so
@@ -125,6 +126,7 @@ celerybeat.pid
 *.sage.py
 
 # Environments
+.myenv
 .env
 .venv
 env/
@@ -163,3 +165,5 @@ cython_debug/
 #  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/
+
+# data/

.gitignore

@@ -0,0 +1,2 @@
+run-test:
+	PYTHONPATH=src python -m pytest -v

Makefile

@@ -4,4 +4,7 @@ pyarrow
 seaborn
 matplotlib
 jupyter
-pytest
+pytest
+pytest-pythonpath
+torch
+torchvision

src/checkpoint/checkpoint.pth renamed to checkpoint/checkpoint.pth

@@ -0,0 +1 @@
+from .setup_data import *

src/checkpoint/checkpoint_notebook.pth renamed to checkpoint/checkpoint_notebook.pth

@@ -0,0 +1,70 @@
+from torchvision import datasets
+from torch.utils.data import Dataset, DataLoader
+from onepiece_classify.transforms import (
+    get_train_transforms,
+    get_valid_transforms,
+    get_test_transforms,
+)
+from pathlib import Path
+
+
+class OnepieceImageDataLoader:
+    def __init__(
+        self,
+        root_path: str, 
+        batch_size: int = 32, 
+        num_workers: int = 0
+    ):
+        self.root_path: Path = Path(root_path)
+        self.train_path: Path = self.root_path.joinpath("train")
+        self.valid_path: Path = self.root_path.joinpath("val")
+        self.test_path: Path = self.root_path.joinpath("test")
+
+        self.batch_size = batch_size
+        self.num_workers = num_workers
+
+        self.trainset = self._build_dataset(mode="train")
+        self.validset = self._build_dataset(mode="valid")
+        self.testset = self._build_dataset(mode="test")
+
+        self.train_loader = self._build_dataloader(mode="train", shuffle=True)
+        self.valid_loader = self._build_dataloader(mode="valid", shuffle=False)
+        self.test_loader = self._build_dataloader(mode="test", shuffle=False)
+
+
+    def _build_dataset(self, mode='train') -> datasets.ImageFolder:
+        dset = None
+        if mode == "train":
+            trans = get_train_transforms()
+            path = self.train_path
+        elif mode == "valid":
+            trans = get_valid_transforms()
+            path = self.valid_path
+        elif mode == "test":
+            trans = get_test_transforms()
+            path = self.test_path
+        else:
+            trans = get_test_transforms()
+            path = self.test_path
+        
+        dset = datasets.ImageFolder(str(path), transform=trans)
+        return dset
+
+    def _build_dataloader(self, mode: str = "train", shuffle: bool = True) -> DataLoader:
+        if mode == "train":
+            dset = self.trainset
+        elif mode == "valid":
+            dset = self.validset
+        elif mode == "test":
+            dset = self.testset
+        else:
+            dset = self.testset
+
+        loader = DataLoader(
+            dset,
+            batch_size=self.batch_size,
+            shuffle=shuffle,
+            pin_memory=True
+        )
+
+        return loader

src/data/test/Ace/212.png_inverted.png renamed to data/test/Ace/212.png_inverted.png

@@ -0,0 +1,2 @@
+from .base import *
+from .recognition import *

src/data/test/Akainu/modified_203.png renamed to data/test/Akainu/modified_203.png

@@ -0,0 +1,24 @@
+from abc import ABC, abstractmethod
+from typing import Optional
+import numpy as np
+from PIL import Image
+import torch
+
+
+class BaseInference(ABC):
+    
+    @abstractmethod
+    def pre_process(self, image: Optional[str|np.ndarray|Image.Image])->torch.Tensor:
+        pass
+    
+    @abstractmethod
+    def forward(self, image_tensor: torch.Tensor) -> torch.Tensor:
+        pass
+    
+    @abstractmethod
+    def post_process(self, output: torch.Tensor) -> str:
+        pass
+    
+    @abstractmethod
+    def predict(self, image: Optional[str|np.ndarray|Image.Image]) -> dict:
+        pass