first update

.gitignore

@@ -0,0 +1,138 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+pip-wheel-metadata/
+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/
+
+# 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
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+.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
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.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/
+.neptune
+
+#dataset
+data
+crop_data
+examples
+#model
+lightning_logs
+ckpts

app.py

@@ -0,0 +1,124 @@
+from typing import Any
+import pytorch_lightning as pl
+from torchvision.models import  efficientnet_v2_s, EfficientNet_V2_S_Weights
+import torch
+from torch import nn
+from torchvision import transforms
+from torch.nn import functional as F
+import yaml
+from yaml.loader import SafeLoader
+from PIL import Image
+import gradio as gr
+import os
+
+class WeedModel(pl.LightningModule):
+    def __init__(self, params):
+        super().__init__()
+        self.params = params
+     
+        model = self.params["model"]
+
+        if(model.lower() == "efficientnet"):
+            if(self.params["pretrained"]): self.base_model = efficientnet_v2_s(weights=EfficientNet_V2_S_Weights.IMAGENET1K_V1)
+            else: self.base_model = efficientnet_v2_s(weights=None)
+            num_ftrs = self.base_model.classifier[-1].in_features
+            self.base_model.classifier[-1] = nn.Linear(num_ftrs, self.params["n_class"])
+
+        else:
+            print("not prepared model yet!!")
+
+    def forward(self, x):
+        embedding = self.base_model(x)
+        return embedding
+
+    def configure_optimizers(self):
+        if(self.params["optimizer"] == "Adam"):
+            optimizer = torch.optim.Adam(self.parameters(), lr=self.params["Lr"])
+        elif(self.params["optimizer"] == "SGD"):
+            optimizer = torch.optim.SGD(self.parameters(), lr=self.params["Lr"])
+        return optimizer
+
+    def training_step(self, train_batch, batch_idx):
+        x = train_batch["image"]
+        y = train_batch["label"]
+
+        y_hat = self(x)
+        loss = F.cross_entropy(y_hat, y)
+        self.log('metrics/batch/train_loss', loss, prog_bar=False)
+
+        preds = F.softmax(y_hat, dim=-1)
+     
+        return loss
+    
+    def validation_step(self, val_batch, batch_idx):
+        
+        x = val_batch["image"]
+        y = val_batch["label"]
+
+        y_hat = self(x)
+        loss = F.cross_entropy(y_hat, y)
+        self.log('metrics/batch/val_loss', loss)
+    
+    def predict_step(self, batch: Any, batch_idx: int=0, dataloader_idx: int = 0) -> Any:
+        y_hat = self(batch)
+        preds = torch.softmax(y_hat, dim=-1).tolist()
+        
+        # preds = torch.argmax(preds, dim=-1)
+        return preds
+        
+
+def predict(image):
+
+    tensor_image = transform(image)
+    outs = model.predict_step(tensor_image.unsqueeze(0))
+    labels = {class_names[k]: float(v) for k, v in enumerate(outs[0][:-1])}
+
+    return labels
+
+
+title = " AISeed AI Application Demo "
+description = "# A Demo of Deep Learning for Weed Classification"
+example_list = [["examples/" + example] for example in os.listdir("examples")]
+
+with open("class_names.txt", "r", encoding='utf-8') as f:
+    class_names = f.read().splitlines()
+    
+with gr.Blocks() as demo:
+    demo.title = title
+    gr.Markdown(description)
+    with gr.Tabs():
+        with gr.TabItem("for Images"):
+            with gr.Row():
+                with gr.Column():
+                    im = gr.Image(type="pil", label="input image")
+                with gr.Column():
+                    label_conv = gr.Label(label="Predictions", num_top_classes=4)
+                    btn = gr.Button(value="predict")
+            btn.click(predict, inputs=im, outputs=[label_conv])
+            gr.Examples(examples=example_list, inputs=[im], outputs=[label_conv])
+        with gr.TabItem("for Webcam"):
+            with gr.Row():
+                with gr.Column():
+                    webcam = gr.Image(type="pil", label="input image", source="webcam")
+                    # capture = gr.Image(type="pil", label="output image")
+                with gr.Column():
+                    label = gr.Label(label="Predictions", num_top_classes=4)
+                
+            webcam.change(predict, inputs=webcam, outputs=[label])
+    
+       
+if __name__ == '__main__':
+    with open('config.yaml') as f:
+        PARAMS = yaml.load(f, Loader=SafeLoader)
+        print(PARAMS)
+    model = WeedModel.load_from_checkpoint("model\epoch=08.ckpt", params=PARAMS).cpu()
+    model.eval()
+
+    transform = transforms.Compose([
+                            transforms.Resize(256),
+                            transforms.CenterCrop(224),
+                            transforms.ToTensor(),
+                            transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]),
+                        ])
+    
+    demo.launch()

class_names.txt

@@ -0,0 +1,39 @@
+목표_개망초
+목표_개쑥갓
+목표_단풍잎돼지풀
+목표_돼지풀
+목표_망초
+목표_미국가막사리
+목표_방가지똥
+목표_별꽃아재비
+목표_붉은서나물
+목표_울산도깨비바늘
+목표_주홍서나물
+목표_큰도꼬마리
+목표_큰망초
+목표_큰방가지똥
+목표_털별꽃아재비
+목표_둥근잎나팔꽃
+목표_둥근잎미국나팔꽃
+목표_둥근잎유홍초
+목표_미국나팔꽃
+목표_별나팔꽃
+목표_애기나팔꽃
+목표_냄새명아주
+목표_양명아주
+목표_좀명아주
+목표_취명아주
+목표_흰명아주
+목표_가는털비름
+목표_가시비름
+목표_개비름
+목표_긴털비름
+목표_민털비름
+목표_청비름
+목표_털비름
+목표_가는미국외풀
+목표_눈개불알풀
+목표_미국외풀
+목표_선개불알풀
+목표_좀개불알풀
+목표_큰개불알풀

config.yaml

@@ -0,0 +1,19 @@
+{
+    #Dataset
+    "train_path": "./train.txt",
+    "test_path": "./test.txt",
+    "val_path": "./val.txt",
+    "n_data": 1,
+
+
+    #Model
+    "model": "EfficientNet", # [Alexnet, VGG, GoogleNet, ResNet, DenseNet, MobileNet, SqueezeNet, ShuffleNet, EfficientNet, SE-ResNet (not available)]
+    "pretrained": True,
+    "n_class": 40,
+
+    #Training
+    "B_sz": 4,
+    "Lr": 0.001,
+    "Epoch": 5,
+    "optimizer": "Adam" #[Adam, SGD]
+  }

model/epoch=08.ckpt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0c21a880dfa8c41e3c1533f5e6ab5a12a2bc562d99db7c3504b042384a54de1d
+size 244046718

requirements.txt

@@ -0,0 +1,3 @@
+pytorch-lightning
+torch==1.8.1
+torchvision==0.9.1