Updata

6 months ago · 2839a55803
parent 19f5c68ac4
commit 2839a55803
28 changed files with 27663 additions and 50713 deletions
--- a/Paddle_ocr/video_pgnet_infer.py
+++ b/Paddle_ocr/video_pgnet_infer.py
@ -1,127 +0,0 @@
 import argparse
 import os
 import cv2
 import numpy as np
 import onnxruntime
 from paddleocr.ppocr.data.imaug.operators import (
    E2EResizeForTest, KeepKeys, NormalizeImage, ToCHWImage
 )
 # from ppocr.data.imaug.operators import (
 #     E2EResizeForTest, KeepKeys, NormalizeImage, ToCHWImage
 # )
 from paddleocr.ppocr.postprocess.pg_postprocess import PGPostProcess
 # from ppocr.postprocess.pg_postprocess import PGPostProcess
 from pgnet.chr_dct import chr_dct_list
 class PGNetPredictor:
    def __init__(self, model_path, cpu=False):
        self.model_path = model_path
        self.dict_path = "ic15_dict.txt"
        if not os.path.exists(self.dict_path):
            with open(self.dict_path, "w") as f:
                f.writelines(chr_dct_list)
        providers = ["CPUExecutionProvider"] if cpu else ["CUDAExecutionProvider"]
        self.sess = onnxruntime.InferenceSession(model_path, providers=providers)
        self.transforms = [
            E2EResizeForTest(max_side_len=768, valid_set="totaltext"),
            NormalizeImage(scale=1/255.0, mean=[0.485,0.456,0.406],
                           std=[0.229,0.224,0.225], order="hwc"),
            ToCHWImage(),
            KeepKeys(keep_keys=["image", "shape"]),
        ]
        self.pgpostprocess = PGPostProcess(
            character_dict_path=self.dict_path,
            valid_set="totaltext",
            score_thresh=0.5,
            mode="fast",
        )
    def preprocess(self, img):
        self.ori_im = img.copy()
        data = {"image": img}
        for transform in self.transforms:
            data = transform(data)
        img, shape_list = data
        return np.expand_dims(img, axis=0), np.expand_dims(shape_list, axis=0)
    def predict(self, img):
        ort_inputs = {self.sess.get_inputs()[0].name: img}
        outputs = self.sess.run(None, ort_inputs)
        return {
            "f_border": outputs[0],
            "f_char": outputs[1],
            "f_direction": outputs[2],
            "f_score": outputs[3],
        }
    def clip_boxes(self, boxes, shape):
        h, w = shape[:2]
        clipped = []
        for box in boxes:
            box[:, 0] = np.clip(box[:, 0], 0, w - 1)
            box[:, 1] = np.clip(box[:, 1], 0, h - 1)
            clipped.append(box)
        return np.array(clipped)
    def postprocess(self, preds, shape_list):
        result = self.pgpostprocess(preds, shape_list)
        pts, texts = result["points"], result["texts"]
        return self.clip_boxes(pts, self.ori_im.shape), texts
    def infer(self, img):
        img_input, shape = self.preprocess(img)
        preds = self.predict(img_input)
        return self.postprocess(preds, shape)
 def draw_results(frame, boxes, texts):
    for box, text in zip(boxes, texts):
        box = box.astype(int).reshape(-1, 1, 2)
        cv2.polylines(frame, [box], True, (255,255,0), 2)
        cv2.putText(frame, text, tuple(box[0][0]), cv2.FONT_HERSHEY_SIMPLEX, 
                    0.7, (0,255,0), 2)
    return frame
 if __name__ == "__main__":
    parser = argparse.ArgumentParser(description="PGNet Video OCR")
    parser.add_argument("--model", type=str, required=True)
    parser.add_argument("--video", type=str, required=True)
    parser.add_argument("--cpu", action="store_true")
    args = parser.parse_args()
    predictor = PGNetPredictor(args.model, args.cpu)
    cap = cv2.VideoCapture(args.video)
    width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
    height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
    fps = cap.get(cv2.CAP_PROP_FPS)
    out_name = os.path.splitext(os.path.basename(args.video))[0] + "_pgnet_output.mp4"
    out_path = os.path.join(os.path.dirname(args.video), out_name)
    writer = cv2.VideoWriter(out_path, cv2.VideoWriter_fourcc(*"mp4v"), fps, (width, height))
    print(f"▶ Processing video... (Output: {out_path})")
    while True:
        ret, frame = cap.read()
        if not ret:
            break
        boxes, texts = predictor.infer(frame)
        frame = draw_results(frame, boxes, texts)
        writer.write(frame)
    cap.release()
    writer.release()
    print("🎉 Done! Video saved:", out_path)
--- a/YOLO_FACE_EXP/dataset.py
+++ b/YOLO_FACE_EXP/dataset.py
@ -1,39 +0,0 @@
 import torch
 from torchvision import datasets, transforms
 from torch.utils.data import DataLoader
 import os
 def get_face_dataloaders(data_dir, batch_size=64, num_workers=4):
    """
    Input: data_dir (e.g., ~/face_exp/datasets)
    Structure assumed: data_dir/CASIA-WebFace/ID/images.jpg
    """
    # 1. Train Transform (Augmentation + Resize to 128x128)
    train_transform = transforms.Compose([
        transforms.Resize((128, 128), interpolation=transforms.InterpolationMode.BICUBIC),
        transforms.RandomHorizontalFlip(p=0.5),
        transforms.ColorJitter(brightness=0.1, contrast=0.1, saturation=0.1),
        transforms.ToTensor(),
        transforms.Normalize(mean=[0.5, 0.5, 0.5], std=[0.5, 0.5, 0.5])
    ])
    # 2. Path Setup
    train_dir = os.path.join(data_dir, 'CASIA-WebFace')
    if not os.path.exists(train_dir):
        raise FileNotFoundError(f"데이터셋 경로를 찾을 수 없습니다: {train_dir}\n'CASIA-WebFace' 폴더가 해당 위치에 있는지 확인해주세요.")
    # 3. Dataset & Loader
    train_dataset = datasets.ImageFolder(root=train_dir, transform=train_transform)
    train_loader = DataLoader(
        train_dataset,
        batch_size=batch_size,
        shuffle=True,
        num_workers=num_workers,
        pin_memory=True,
        drop_last=True 
    )
    return train_loader, len(train_dataset.classes)
--- a/YOLO_FACE_EXP/export_onnx.py
+++ b/YOLO_FACE_EXP/export_onnx.py
@ -1,71 +0,0 @@
 import torch
 import argparse
 import os
 from model import get_face_model
 def export_to_onnx(pt_path, output_path):
    print(f"🔄 Loading model from: {pt_path}")
    # 1. 모델 구조 불러오기 및 가중치 로드
    # 보드용 모델이므로 학습용 ArcFace 헤더는 버리고, Backbone만 가져옵니다.
    model = get_face_model()
    # CPU로 로드 (변환은 굳이 GPU 불필요)
    checkpoint = torch.load(pt_path, map_location='cpu')
    # state_dict 로드 (혹시 모를 키 불일치 방지를 위해 strict=False는 선택사항이나, 여기선 구조가 같으므로 True 권장)
    try:
        model.load_state_dict(checkpoint)
    except RuntimeError as e:
        print(f"⚠️ Key mismatch detected. Trying to load with strict=False...")
        model.load_state_dict(checkpoint, strict=False)
    # 2. Eval 모드 전환 (매우 중요)
    # 이걸 안 하면 BatchNorm, Dropout 등이pip  학습 모드로 동작하여 결과가 이상해집니다.
    model.eval()
    # 3. Dummy Input 생성 (Static Shape: 128x128)
    # 보드 사양에 맞춰 배치 사이즈는 1로 고정합니다. [1, 3, 128, 128]
    dummy_input = torch.randn(1, 3, 128, 128)
    print(f"Target ONNX Path: {output_path}")
    # 4. ONNX Export
    # external_data=False는 PyTorch export에서 기본적으로 2GB 미만 모델에 대해 적용되어 단일 파일로 나옵니다.
    # dynamic_axes 옵션을 뺌으로써 Static Shape을 강제합니다.
    torch.onnx.export(
        model,                      # 실행될 모델
        dummy_input,                # 모델 입력값 (차원 체크용)
        output_path,                # 저장될 경로
        # export_params=True,         # 모델 파일 안에 웨이트 저장 (external_data=False 효과)
        # opset_version=11,           # 임베디드 보드에서 가장 호환성 좋은 버전 (11 추천)
        # do_constant_folding=True,   # 상수 폴딩 최적화
        input_names=['input'],      # 입력 노드 이름
        output_names=['output'],     # 출력 노드 이름
        external_data=False
        # dynamic_axes={...}  <-- 이 옵션을 사용하지 않음으로써 Static Shape으로 고정됨!
    )
    print(f"✅ Conversion Completed! Model saved at: {output_path}")
    print(f"ℹ️  Input Shape: {dummy_input.shape} (Static)")
    print(f"ℹ️  Please check if '{output_path}' is a single file.")
 if __name__ == "__main__":
    parser = argparse.ArgumentParser(description='Convert PyTorch model to ONNX')
    # 입력받을 .pt 파일 경로
    parser.add_argument('--input', type=str, required=True, help='Input .pt file path')
    # 출력할 .onnx 파일 경로 (옵션)
    parser.add_argument('--output', type=str, default=None, help='Output .onnx file path')
    args = parser.parse_args()
    # Output 경로가 없으면 Input 경로에서 확장자만 바꿔서 자동 지정
    if args.output is None:
        args.output = args.input.replace('.pt', '.onnx')
    if not os.path.exists(args.input):
        print(f"❌ Error: Input file not found: {args.input}")
    else:
        export_to_onnx(args.input, args.output)
--- a/YOLO_FACE_EXP/face_recog_exp.ipynb
+++ b/YOLO_FACE_EXP/face_recog_exp.ipynb
--- a/YOLO_FACE_EXP/model.py
+++ b/YOLO_FACE_EXP/model.py
@ -1,149 +0,0 @@
 import torch
 import torch.nn as nn
 import torch.nn.functional as F
 import math
 # --------------------------------------------------------
 # Basic Blocks (Quantization Friendly: ReLU used)
 # --------------------------------------------------------
 def conv3x3(in_planes, out_planes, stride=1):
    return nn.Conv2d(in_planes, out_planes, kernel_size=3, stride=stride,
                     padding=1, bias=False)
 class Bottleneck(nn.Module):
    expansion = 4
    def __init__(self, inplanes, planes, stride=1, downsample=None):
        super(Bottleneck, self).__init__()
        # Kernel size restricted to 1 and 3 for board compatibility
        self.conv1 = nn.Conv2d(inplanes, planes, kernel_size=1, bias=False)
        self.bn1 = nn.BatchNorm2d(planes)
        self.conv2 = nn.Conv2d(planes, planes, kernel_size=3, stride=stride, padding=1, bias=False)
        self.bn2 = nn.BatchNorm2d(planes)
        self.conv3 = nn.Conv2d(planes, planes * self.expansion, kernel_size=1, bias=False)
        self.bn3 = nn.BatchNorm2d(planes * self.expansion)
        # self.relu = nn.ReLU(inplace=True) # PReLU -> ReLU
        self.relu = nn.ReLU(inplace=False) # PReLU -> ReLU
        self.downsample = downsample
        self.stride = stride
    def forward(self, x):
        residual = x
        out = self.conv1(x)
        out = self.bn1(out)
        out = self.relu(out)
        out = self.conv2(out)
        out = self.bn2(out)
        out = self.relu(out)
        out = self.conv3(out)
        out = self.bn3(out)
        if self.downsample is not None:
            residual = self.downsample(x)
        out += residual
        out = self.relu(out)
        return out
 # --------------------------------------------------------
 # Backbone: Modified ResNet-50
 # --------------------------------------------------------
 class ResNetFace(nn.Module):
    def __init__(self, block, layers, embedding_size=128):
        super(ResNetFace, self).__init__()
        self.inplanes = 64
        # Stem Block: 7x7 replaced by three 3x3 convs
        self.stem = nn.Sequential(
            nn.Conv2d(3, 64, kernel_size=3, stride=1, padding=1, bias=False),
            nn.BatchNorm2d(64),
            # nn.ReLU(inplace=True),
            nn.ReLU(inplace=False),
            nn.Conv2d(64, 64, kernel_size=3, stride=1, padding=1, bias=False),
            nn.BatchNorm2d(64),
            # nn.ReLU(inplace=True),
            nn.ReLU(inplace=False),
            nn.Conv2d(64, 64, kernel_size=3, stride=2, padding=1, bias=False),
            nn.BatchNorm2d(64),
            # nn.ReLU(inplace=True),
            nn.ReLU(inplace=False),
        )
        self.maxpool = nn.MaxPool2d(kernel_size=3, stride=2, padding=1)
        self.layer1 = self._make_layer(block, 64, layers[0])
        self.layer2 = self._make_layer(block, 128, layers[1], stride=2)
        self.layer3 = self._make_layer(block, 256, layers[2], stride=2)
        self.layer4 = self._make_layer(block, 512, layers[3], stride=2)
        # Output Head: 128x128 Input -> 4x4 Feature Map
        self.avgpool = nn.AvgPool2d(4) # Result: 1x1
        # FC layer replaced by 1x1 Conv for 128-d embedding
        self.fc_conv = nn.Conv2d(512 * block.expansion, embedding_size, kernel_size=1, bias=False)
        self.bn_last = nn.BatchNorm2d(embedding_size)
        # Init weights
        for m in self.modules():
            if isinstance(m, nn.Conv2d):
                nn.init.kaiming_normal_(m.weight, mode='fan_out', nonlinearity='relu')
            elif isinstance(m, nn.BatchNorm2d):
                nn.init.constant_(m.weight, 1)
                nn.init.constant_(m.bias, 0)
    def _make_layer(self, block, planes, blocks, stride=1):
        downsample = None
        if stride != 1 or self.inplanes != planes * block.expansion:
            downsample = nn.Sequential(
                nn.Conv2d(self.inplanes, planes * block.expansion,
                          kernel_size=1, stride=stride, bias=False),
                nn.BatchNorm2d(planes * block.expansion),
            )
        layers = []
        layers.append(block(self.inplanes, planes, stride, downsample))
        self.inplanes = planes * block.expansion
        for _ in range(1, blocks):
            layers.append(block(self.inplanes, planes))
        return nn.Sequential(*layers)
    def forward(self, x):
        x = self.stem(x)
        x = self.maxpool(x)
        x = self.layer1(x)
        x = self.layer2(x)
        x = self.layer3(x)
        x = self.layer4(x)
        x = self.avgpool(x)
        x = self.fc_conv(x)
        x = self.bn_last(x) # Output: [N, 128, 1, 1]
        return x
 # --------------------------------------------------------
 # ArcFace Loss Header
 # --------------------------------------------------------
 class ArcFace(nn.Module):
    def __init__(self, in_features, out_features, s=64.0, m=0.50):
        super(ArcFace, self).__init__()
        self.in_features = in_features
        self.out_features = out_features
        self.s = s
        self.m = m
        self.weight = nn.Parameter(torch.FloatTensor(out_features, in_features))
        nn.init.xavier_uniform_(self.weight)
    def forward(self, input, label):
        # Flatten [N, 128, 1, 1] -> [N, 128] for loss calculation
        embedding = input.view(input.size(0), -1)
        cosine = F.linear(F.normalize(embedding), F.normalize(self.weight))
        # Stable implementation of ArcFace
        theta = torch.acos(torch.clamp(cosine, -1.0 + 1e-7, 1.0 - 1e-7))
        target_logits = torch.cos(theta + self.m)
        one_hot = torch.zeros_like(cosine)
        one_hot.scatter_(1, label.view(-1, 1).long(), 1)
        output = one_hot * target_logits + (1.0 - one_hot) * cosine
        output *= self.s
        return output
 def get_face_model():
    return ResNetFace(Bottleneck, [3, 4, 6, 3], embedding_size=128)
--- a/YOLO_FACE_EXP/train_ver1.py
+++ b/YOLO_FACE_EXP/train_ver1.py
@ -1,120 +0,0 @@
 import torch
 import torch.nn as nn
 import torch.optim as optim
 import os
 import argparse
 from tqdm import tqdm
 from datetime import datetime  # 시간 정보를 가져오기 위해 추가
 # Import our custom modules
 from model import get_face_model, ArcFace
 from dataset import get_face_dataloaders
 def main():
    # --------------------------------------------------------
    # 1. Hyperparameters & Settings
    # --------------------------------------------------------
    parser = argparse.ArgumentParser(description='Face Recognition Training for Apache 6')
    # 데이터셋 경로 (사용자 환경에 맞게 기본값 설정)
    parser.add_argument('--data_dir', type=str, default='/home/cuuva/face_exp/datasets', help='Path to datasets')
    # 결과 저장 최상위 경로 (여기 아래에 시간별 폴더가 생김)
    parser.add_argument('--project_dir', type=str, default='./results', help='Base directory for results')
    parser.add_argument('--epochs', type=int, default=20, help='Number of epochs')
    parser.add_argument('--batch_size', type=int, default=64, help='Batch size')
    parser.add_argument('--lr', type=float, default=0.1, help='Learning rate')
    args = parser.parse_args()
    device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
    print(f"Device: {device}")
    # --------------------------------------------------------
    # [수정됨] 실험 디렉토리 생성 로직 (yy-mm-dd-hour-minute)
    # --------------------------------------------------------
    # 현재 시간 구하기
    current_time = datetime.now().strftime("%y-%m-%d-%H-%M")
    # 최종 저장 경로: ./results/23-12-12-15-30/
    save_dir = os.path.join(args.project_dir, current_time)
    # 폴더 생성
    os.makedirs(save_dir, exist_ok=True)
    print(f"✅ Experiment results will be saved to: {save_dir}")
    # --------------------------------------------------------
    # 2. Load Data
    # --------------------------------------------------------
    print("Loading Data...")
    try:
        train_loader, num_classes = get_face_dataloaders(args.data_dir, args.batch_size)
        print(f"Classes (People): {num_classes}, Batch Size: {args.batch_size}")
    except Exception as e:
        print(f"❌ Error loading data: {e}")
        return
    # --------------------------------------------------------
    # 3. Initialize Model & Loss
    # --------------------------------------------------------
    # Backbone (보드에 배포할 모델)
    backbone = get_face_model().to(device)
    # ArcFace Header (학습용 Loss 계산기)
    metric_fc = ArcFace(in_features=128, out_features=num_classes).to(device)
    # Loss
    criterion = nn.CrossEntropyLoss()
    # Optimizer
    optimizer = optim.SGD([
        {'params': backbone.parameters()},
        {'params': metric_fc.parameters()}
    ], lr=args.lr, momentum=0.9, weight_decay=5e-4)
    # Scheduler
    scheduler = optim.lr_scheduler.MultiStepLR(optimizer, milestones=[8, 14, 18], gamma=0.1)
    # --------------------------------------------------------
    # 4. Training Loop
    # --------------------------------------------------------
    print("🚀 Start Training...")
    for epoch in range(args.epochs):
        backbone.train()
        metric_fc.train()
        running_loss = 0.0
        pbar = tqdm(train_loader, desc=f"Epoch {epoch+1}/{args.epochs}")
        for images, labels in pbar:
            images, labels = images.to(device), labels.to(device)
            optimizer.zero_grad()
            # Forward Pass
            features = backbone(images)       # [N, 128, 1, 1]
            outputs = metric_fc(features, labels) # [N, Num_Classes]
            # Loss Calc & Backward
            loss = criterion(outputs, labels)
            loss.backward()
            optimizer.step()
            running_loss += loss.item()
            pbar.set_postfix({'loss': running_loss / (pbar.n + 1)})
        scheduler.step()
        # [수정됨] Save Checkpoint (.pt 확장자 사용)
        # 해당 실험 폴더(save_dir) 안에 저장됨
        save_path = os.path.join(save_dir, f"backbone_epoch_{epoch+1}.pt")
        torch.save(backbone.state_dict(), save_path)
        # 마지막 에폭일 때 로그 출력
        if epoch == args.epochs - 1:
            print(f"🎉 Training Finished! Final model saved at: {save_path}")
 if __name__ == "__main__":
    main()
--- a/YOLO_FACE_EXP/train_ver2.py
+++ b/YOLO_FACE_EXP/train_ver2.py
@ -1,141 +0,0 @@
 import torch
 import torch.nn as nn
 import torch.optim as optim
 import os
 import argparse
 from tqdm import tqdm
 from datetime import datetime
 from torchvision import transforms
 from torch.utils.data import DataLoader
 # Import Custom Modules
 from model import get_face_model, ArcFace
 from dataset import get_face_dataloaders
 from validation import LFWDataset, validate_lfw
 def main():
    # 1. Hyperparameters
    parser = argparse.ArgumentParser(description='Face Recognition Training with LFW Validation')
    parser.add_argument('--data_dir', type=str, default='/home/cuuva/face_exp/datasets', help='Root dataset dir')
    parser.add_argument('--project_dir', type=str, default='./results', help='Save dir')
    parser.add_argument('--epochs', type=int, default=20)
    parser.add_argument('--batch_size', type=int, default=64)
    parser.add_argument('--lr', type=float, default=0.1)
    args = parser.parse_args()
    device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
    print(f"Device: {device}")
    # 디렉토리 설정
    current_time = datetime.now().strftime("%y-%m-%d-%H-%M")
    base_save_dir = os.path.join(args.project_dir, current_time)
    best_save_dir = os.path.join(base_save_dir, 'best_model')
    os.makedirs(best_save_dir, exist_ok=True)
    print(f"✅ Save path: {base_save_dir}")
    # 2. Train Data Loader
    print("Loading Train Data...")
    try:
        train_loader, num_classes = get_face_dataloaders(args.data_dir, args.batch_size)
    except Exception as e:
        print(f"❌ Train Data Error: {e}")
        return
    # --------------------------------------------------------
    # [수정됨] 3. LFW Validation Loader (경로 로직 개선)
    # --------------------------------------------------------
    print("Loading LFW Data...")
    # pairs.txt가 있는 루트 경로: ~/face_exp/datasets/LFW
    lfw_root = os.path.join(args.data_dir, 'LFW')
    pairs_path = os.path.join(lfw_root, 'pairs.txt') 
    # 실제 이미지가 있는 경로 찾기
    # 1순위: ~/face_exp/datasets/LFW/lfw (보여주신 구조)
    # 2순위: ~/face_exp/datasets/LFW (일반적인 구조)
    if os.path.exists(os.path.join(lfw_root, 'lfw')):
        lfw_img_dir = os.path.join(lfw_root, 'lfw')
    else:
        lfw_img_dir = lfw_root
    print(f"ℹ️  Pairs path: {pairs_path}")
    print(f"ℹ️  Images path: {lfw_img_dir}")
    if os.path.exists(pairs_path) and os.path.exists(lfw_img_dir):
        lfw_transform = transforms.Compose([
            transforms.Resize((128, 128), interpolation=transforms.InterpolationMode.BICUBIC),
            transforms.ToTensor(),
            transforms.Normalize(mean=[0.5, 0.5, 0.5], std=[0.5, 0.5, 0.5])
        ])
        # 여기서 lfw_img_dir를 넘겨주는 것이 핵심!
        lfw_dataset = LFWDataset(lfw_img_dir, pairs_path, transform=lfw_transform)
        lfw_loader = DataLoader(lfw_dataset, batch_size=64, shuffle=False, num_workers=4, drop_last=False)
        do_validation = True
        print(f"✅ LFW Loaded. Pairs: {len(lfw_dataset)}")
    else:
        print(f"⚠️ Warning: 'pairs.txt' or Image dir not found. Skipping Validation.")
        do_validation = False
    # 4. Model & Loss
    backbone = get_face_model().to(device)
    metric_fc = ArcFace(in_features=128, out_features=num_classes).to(device)
    criterion = nn.CrossEntropyLoss()
    optimizer = optim.SGD([
        {'params': backbone.parameters()},
        {'params': metric_fc.parameters()}
    ], lr=args.lr, momentum=0.9, weight_decay=5e-4)
    scheduler = optim.lr_scheduler.MultiStepLR(optimizer, milestones=[8, 14, 18], gamma=0.1)
    # 5. Training Loop
    print("🚀 Start Training...")
    best_acc = 0.0 
    for epoch in range(args.epochs):
        # --- TRAIN ---
        backbone.train()
        metric_fc.train()
        running_loss = 0.0
        pbar = tqdm(train_loader, desc=f"Epoch {epoch+1}/{args.epochs}")
        for images, labels in pbar:
            images, labels = images.to(device), labels.to(device)
            optimizer.zero_grad()
            features = backbone(images)
            outputs = metric_fc(features, labels)
            loss = criterion(outputs, labels)
            loss.backward()
            optimizer.step()
            running_loss += loss.item()
            pbar.set_postfix({'loss': f"{running_loss / (pbar.n + 1):.4f}"})
        scheduler.step()
        # --- VALIDATION (LFW) ---
        if do_validation:
            acc, th = validate_lfw(backbone, lfw_loader, device)
            print(f"📊 LFW Acc: {acc*100:.2f}% (Threshold: {th:.2f})")
            if acc > best_acc:
                best_acc = acc
                save_path = os.path.join(best_save_dir, "best_backbone.pt")
                torch.save(backbone.state_dict(), save_path)
                print(f"🏆 Best Model Updated! Saved to {save_path}")
        else:
            save_path = os.path.join(base_save_dir, f"backbone_epoch_{epoch+1}.pt")
            torch.save(backbone.state_dict(), save_path)
        torch.save(backbone.state_dict(), os.path.join(base_save_dir, "last_backbone.pt"))
    print("\n🎉 Training Finished!")
 if __name__ == "__main__":
    main()
--- a/YOLO_FACE_EXP/validation.py
+++ b/YOLO_FACE_EXP/validation.py
@ -1,119 +0,0 @@
 import torch
 import torch.nn.functional as F
 from torch.utils.data import Dataset, DataLoader
 from torchvision import transforms
 from PIL import Image
 import os
 import numpy as np
 # --------------------------------------------------------
 # 1. LFW Dataset Loader
 # --------------------------------------------------------
 class LFWDataset(Dataset):
    def __init__(self, lfw_dir, pairs_path, transform=None):
        self.lfw_dir = lfw_dir
        self.pairs_path = pairs_path
        self.transform = transform
        self.validation_images = self.get_lfw_paths(lfw_dir)
    def get_lfw_paths(self, lfw_dir):
        # pairs.txt 파싱하여 이미지 경로 쌍과 정답(issame) 리스트 생성
        pairs = []
        with open(self.pairs_path, 'r') as f:
            lines = f.readlines()[1:] # 첫 줄(헤더) 건너뜀
        for line in lines:
            p = line.strip().split('\t')
            if len(p) == 3: # 같은 사람 (name, img1_num, img2_num)
                name = p[0]
                img1 = os.path.join(lfw_dir, name, f"{name}_{int(p[1]):04d}.jpg")
                img2 = os.path.join(lfw_dir, name, f"{name}_{int(p[2]):04d}.jpg")
                issame = True
                pairs.append((img1, img2, issame))
            elif len(p) == 4: # 다른 사람 (name1, img1_num, name2, img2_num)
                name1 = p[0]
                img1 = os.path.join(lfw_dir, name1, f"{name1}_{int(p[1]):04d}.jpg")
                name2 = p[2]
                img2 = os.path.join(lfw_dir, name2, f"{name2}_{int(p[3]):04d}.jpg")
                issame = False
                pairs.append((img1, img2, issame))
        return pairs
    def __len__(self):
        return len(self.validation_images)
    def __getitem__(self, index):
        img1_path, img2_path, issame = self.validation_images[index]
        try:
            img1 = Image.open(img1_path).convert('RGB')
            img2 = Image.open(img2_path).convert('RGB')
        except Exception as e:
            # 혹시 파일이 없을 경우를 대비한 더미 (실제론 파일 확인 필요)
            print(f"File Load Error: {e}")
            img1 = Image.new('RGB', (128, 128))
            img2 = Image.new('RGB', (128, 128))
        if self.transform:
            img1 = self.transform(img1)
            img2 = self.transform(img2)
        return img1, img2, issame
 # --------------------------------------------------------
 # 2. Evaluation Function
 # --------------------------------------------------------
 def validate_lfw(model, lfw_loader, device):
    model.eval()
    similarities = []
    actual_issame = []
    print("🔍 Validating on LFW...")
    with torch.no_grad():
        for img1, img2, issame in lfw_loader:
            img1, img2 = img1.to(device), img2.to(device)
            # Feature Extraction
            feat1 = model(img1) # [B, 128, 1, 1]
            feat2 = model(img2) # [B, 128, 1, 1]
            # Flatten
            feat1 = feat1.view(feat1.size(0), -1)
            feat2 = feat2.view(feat2.size(0), -1)
            # Cosine Similarity Calculation
            # 128차원 벡터의 코사인 유사도 (-1 ~ 1)
            cos_sim = F.cosine_similarity(feat1, feat2)
            similarities.extend(cos_sim.cpu().numpy())
            actual_issame.extend(issame.numpy())
    similarities = np.array(similarities)
    actual_issame = np.array(actual_issame)
    # ----------------------------------------------------
    # Best Threshold Search (단순화된 버전)
    # ----------------------------------------------------
    best_acc = 0.0
    best_th = 0.0
    # -1.0 부터 1.0 까지 0.01 단위로 Threshold를 이동하며 정확도 측정
    thresholds = np.arange(-1.0, 1.0, 0.01)
    for th in thresholds:
        # th보다 크면 True(동일인), 작으면 False(타인) 예측
        predict_issame = np.greater(similarities, th)
        # 정답과 비교
        true_positives = np.sum(np.logical_and(predict_issame, actual_issame))
        true_negatives = np.sum(np.logical_and(np.logical_not(predict_issame), np.logical_not(actual_issame)))
        acc = (true_positives + true_negatives) / len(actual_issame)
        if acc > best_acc:
            best_acc = acc
            best_th = th
    return best_acc, best_th
--- a/custom_LP_detect/custom_LP.yaml
+++ b/custom_LP_detect/custom_LP.yaml
@ -1,26 +1,3 @@
 # # path: datasets/VisDrone  # 데이터 경로
 # train: /home/cuuva/aihub_car/CCTV_car_or_licenseplate/data/Training/image/license_plate/
 # val: /home/cuuva/aihub_car/CCTV_car_or_licenseplate/data/Validation/image/license_plate/
 # # test: images/test
 # # path: /home/cuuva/aihub_car/CCTV_car_or_licenseplate/data/Training/
 # # train:
 # #   - image/license_plate/        # 이미지 폴더
 # # val:
 # #   - ../Validation/image/license_plate/
 # # 각 이미지에 대응하는 라벨 경로를 직접 지정
 # labels:
 #   train: /home/cuuva/aihub_car/CCTV_car_or_licenseplate/data/Training/labels/license_plate/
 #   val: /home/cuuva/aihub_car/CCTV_car_or_licenseplate/data/Validation/labels/license_plate/
 # nc: 1
 # names: ['lp']
 # custom_LP.yaml
 # 데이터셋 경로
 train: /home/cuuva/aihub_car/CCTV_car_or_licenseplate/data/Training/images/license_plate/
 val: /home/cuuva/aihub_car/CCTV_car_or_licenseplate/data/Validation/images/license_plate/
--- a/custom_LP_detect/easyocr.py
+++ b/custom_LP_detect/easyocr.py
@ -1,38 +0,0 @@
 import easyocr
 import cv2
 from matplotlib import pyplot as plt
 import time  # ⬅ 추가
 # 1. EasyOCR Reader 생성
 reader = easyocr.Reader(['ko', 'en'], gpu=False)
 # 2. 이미지 불러오기
 # image_path = '/home/cuuva/다운로드/test/ocr_resized.png'
 image_path = '/home/cuuva/experiment/custom_LP_detect/ocr2.png'
 image = cv2.imread(image_path)
 image_rgb = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
 # ------------------------
 # 3. OCR 수행 및 시간 측정
 # ------------------------
 start_time = time.time()
 results = reader.readtext(image_rgb, detail=1)
 end_time = time.time()
 print(f"Inference time: {end_time - start_time:.3f} seconds")
 # 4. 결과 출력 및 시각화
 for (bbox, text, prob) in results:
    print(f"Detected text: {text}, Confidence: {prob:.2f}")
    # 바운딩 박스
    top_left = tuple(map(int, bbox[0]))
    bottom_right = tuple(map(int, bbox[2]))
    cv2.rectangle(image_rgb, top_left, bottom_right, (0, 255, 0), 2)
    cv2.putText(image_rgb, text, (top_left[0], top_left[1]-10), 
                cv2.FONT_HERSHEY_SIMPLEX, 0.8, (0, 0, 255), 2)
 # 5. 시각화
 plt.figure(figsize=(10,6))
 plt.imshow(image_rgb)
 plt.axis('off')
 plt.show()
--- a/fashion_yolo/aihub_exp.ipynb
+++ b/fashion_yolo/aihub_exp.ipynb
--- a/fashion_yolo/convert_labels.py
+++ b/fashion_yolo/convert_labels.py
@ -0,0 +1,94 @@
 import os
 import glob
 from tqdm.notebook import tqdm  # 주피터 노트북용 진행바
 # ==========================================
 # 1. 절대 경로 설정 (수정된 부분)
 # ==========================================
 # 사용자 홈 디렉토리(/home/cuuva)를 포함한 전체 경로를 입력합니다.
 SRC_ROOT = '/home/cuuva/git/Detection_Experiment/datasets/fashionpedia_yolo/labels_all_bak'
 DST_ROOT = '/home/cuuva/git/Detection_Experiment/datasets/fashionpedia_yolo/labels_reduced'
 # ==========================================
 # 2. 클래스 매핑 규칙 (Old ID -> New ID)
 # ==========================================
 # 0(shirt), 1(top), 2(sweater) -> 0 (shirt)
 # 3(cardigan), 4(jacket)       -> 1 (jacket)
 # 6(pants)                     -> 2 (pants)
 # 13(glasses)                  -> 3 (glasses)
 class_mapping = {
    0: 0, 1: 0, 2: 0,
    3: 1, 4: 1,
    6: 2,
    13: 3
 }
 print(f"원본 경로: {SRC_ROOT}")
 print(f"저장 경로: {DST_ROOT}")
 print("-" * 30)
 # ==========================================
 # 3. 데이터 변환 로직
 # ==========================================
 def process_yolo_labels(src_root, dst_root, mapping):
    subsets = ['train', 'val']
    total_files = 0
    total_objects_kept = 0
    for subset in subsets:
        src_dir = os.path.join(src_root, subset)
        dst_dir = os.path.join(dst_root, subset)
        # 소스 디렉토리 존재 확인
        if not os.path.exists(src_dir):
            print(f"⚠️ 에러: 소스 폴더를 찾을 수 없습니다 -> {src_dir}")
            continue
        # 타겟 디렉토리 생성
        os.makedirs(dst_dir, exist_ok=True)
        # 파일 목록 로드
        txt_files = glob.glob(os.path.join(src_dir, '*.txt'))
        total_files += len(txt_files)
        print(f"🚀 Processing [{subset}]: {len(txt_files)} files found.")
        # 변환 시작
        for file_path in tqdm(txt_files, desc=f"{subset} Converting"):
            file_name = os.path.basename(file_path)
            dst_path = os.path.join(dst_dir, file_name)
            new_lines = []
            with open(file_path, 'r') as f:
                lines = f.readlines()
                for line in lines:
                    parts = line.strip().split()
                    if not parts: continue
                    old_cls = int(parts[0])
                    coords = parts[1:] # x, y, w, h
                    # 매핑 규칙에 있는 클래스만 남김
                    if old_cls in mapping:
                        new_cls = mapping[old_cls]
                        new_line = f"{new_cls} {' '.join(coords)}\n"
                        new_lines.append(new_line)
                        total_objects_kept += 1
            # 파일 쓰기 (빈 파일이라도 생성하여 구조 유지)
            with open(dst_path, 'w') as f:
                f.writelines(new_lines)
    return total_files, total_objects_kept
 # 실행
 processed_cnt, kept_cnt = process_yolo_labels(SRC_ROOT, DST_ROOT, class_mapping)
 print("="*30)
 print("✅ 변환 완료")
 print(f"총 처리된 파일: {processed_cnt}개")
 print(f"남은 객체 수: {kept_cnt}개")
 print(f"저장된 위치: {DST_ROOT}")
--- a/fashion_yolo/fashionpedia_exp/yolov8m_fashion_4class/args.yaml
+++ b/fashion_yolo/fashionpedia_exp/yolov8m_fashion_4class/args.yaml
@ -0,0 +1,106 @@
 task: detect
 mode: train
 model: yolov8m.pt
 data: /home/cuuva/git/Detection_Experiment/datasets/fashionpedia_yolo/fashionpedia_custom.yaml
 epochs: 500
 time: null
 patience: 50
 batch: -1
 imgsz: 640
 save: true
 save_period: -1
 cache: false
 device: '0'
 workers: 8
 project: fashionpedia_exp
 name: yolov8m_fashion_4class
 exist_ok: false
 pretrained: true
 optimizer: AdamW
 verbose: true
 seed: 0
 deterministic: true
 single_cls: false
 rect: false
 cos_lr: false
 close_mosaic: 10
 resume: false
 amp: true
 fraction: 1.0
 profile: false
 freeze: null
 multi_scale: false
 compile: false
 overlap_mask: true
 mask_ratio: 4
 dropout: 0.0
 val: true
 split: val
 save_json: false
 conf: null
 iou: 0.7
 max_det: 300
 half: false
 dnn: false
 plots: true
 source: null
 vid_stride: 1
 stream_buffer: false
 visualize: false
 augment: false
 agnostic_nms: false
 classes: null
 retina_masks: false
 embed: null
 show: false
 save_frames: false
 save_txt: false
 save_conf: false
 save_crop: false
 show_labels: true
 show_conf: true
 show_boxes: true
 line_width: null
 format: torchscript
 keras: false
 optimize: false
 int8: false
 dynamic: false
 simplify: true
 opset: null
 workspace: null
 nms: false
 lr0: 0.001
 lrf: 0.01
 momentum: 0.937
 weight_decay: 0.0005
 warmup_epochs: 3.0
 warmup_momentum: 0.8
 warmup_bias_lr: 0.1
 box: 7.5
 cls: 0.5
 dfl: 1.5
 pose: 12.0
 kobj: 1.0
 nbs: 64
 hsv_h: 0.015
 hsv_s: 0.7
 hsv_v: 0.4
 degrees: 0.0
 translate: 0.1
 scale: 0.5
 shear: 0.0
 perspective: 0.0
 flipud: 0.0
 fliplr: 0.5
 bgr: 0.0
 mosaic: 1.0
 mixup: 0.0
 cutmix: 0.0
 copy_paste: 0.0
 copy_paste_mode: flip
 auto_augment: randaugment
 erasing: 0.4
 cfg: null
 tracker: botsort.yaml
 save_dir: /home/cuuva/git/Detection_Experiment/fashion_yolo/fashionpedia_exp/yolov8m_fashion_4class
--- a/fashion_yolo/fashionpedia_exp/yolov8m_fashion_4class/results.csv
+++ b/fashion_yolo/fashionpedia_exp/yolov8m_fashion_4class/results.csv
@ -0,0 +1,119 @@
 epoch,time,train/box_loss,train/cls_loss,train/dfl_loss,metrics/precision(B),metrics/recall(B),metrics/mAP50(B),metrics/mAP50-95(B),val/box_loss,val/cls_loss,val/dfl_loss,lr/pg0,lr/pg1,lr/pg2
 1,474.596,0.93903,1.41545,1.28627,0.61587,0.56472,0.58086,0.39693,1.0572,1.75433,1.3424,0.067021,0.000333122,0.000333122
 2,945.472,0.92099,1.26934,1.27601,0.65767,0.66997,0.70237,0.52208,0.87901,1.11222,1.22253,0.0340196,0.000665135,0.000665135
 3,1415.01,0.88692,1.19259,1.25441,0.69552,0.68564,0.72146,0.53996,0.85246,1.07178,1.18507,0.00101701,0.000995829,0.000995829
 4,1884.52,0.83982,1.12034,1.21812,0.7636,0.69239,0.76012,0.59898,0.75789,0.91664,1.14181,0.00099406,0.00099406,0.00099406
 5,2354.23,0.79317,1.04793,1.18944,0.77686,0.72369,0.79517,0.63704,0.70919,0.82991,1.09923,0.00099208,0.00099208,0.00099208
 6,2823.81,0.76062,0.99744,1.16595,0.79611,0.75397,0.81301,0.66144,0.68034,0.79346,1.09095,0.0009901,0.0009901,0.0009901
 7,3293.38,0.73424,0.95944,1.14808,0.77811,0.77505,0.82129,0.66934,0.6387,0.76163,1.05014,0.00098812,0.00098812,0.00098812
 8,3762.86,0.72051,0.93364,1.14222,0.8307,0.75062,0.83974,0.68937,0.62044,0.70965,1.05224,0.00098614,0.00098614,0.00098614
 9,4232.56,0.69741,0.89999,1.12531,0.80277,0.78105,0.84217,0.7019,0.58509,0.68064,1.02066,0.00098416,0.00098416,0.00098416
 10,4701.9,0.68169,0.87808,1.11663,0.82917,0.79387,0.85449,0.71584,0.58173,0.65823,1.02224,0.00098218,0.00098218,0.00098218
 11,5171.66,0.67499,0.86255,1.11537,0.84073,0.77441,0.86241,0.72051,0.57155,0.65323,1.02392,0.0009802,0.0009802,0.0009802
 12,5641.02,0.66595,0.845,1.10818,0.83944,0.78124,0.86142,0.72778,0.55057,0.63859,1.00338,0.00097822,0.00097822,0.00097822
 13,6110.43,0.65533,0.82281,1.0977,0.85016,0.7874,0.86618,0.73724,0.54842,0.62593,1.00401,0.00097624,0.00097624,0.00097624
 14,6579.99,0.64797,0.81258,1.09286,0.83547,0.80285,0.87085,0.7376,0.53801,0.61695,0.99101,0.00097426,0.00097426,0.00097426
 15,7049.65,0.63715,0.7988,1.08595,0.85961,0.79098,0.87703,0.74884,0.52471,0.59811,0.98177,0.00097228,0.00097228,0.00097228
 16,7519.23,0.62896,0.78492,1.07928,0.82386,0.83591,0.87836,0.75043,0.52344,0.58876,0.98531,0.0009703,0.0009703,0.0009703
 17,7988.97,0.62132,0.77465,1.07472,0.86651,0.79318,0.87996,0.75302,0.51632,0.58756,0.981,0.00096832,0.00096832,0.00096832
 18,8458.46,0.61949,0.76321,1.07346,0.83482,0.82578,0.88359,0.76031,0.51302,0.57436,0.97926,0.00096634,0.00096634,0.00096634
 19,8927.76,0.61162,0.75548,1.07409,0.82063,0.84553,0.8846,0.76373,0.5105,0.57243,0.97865,0.00096436,0.00096436,0.00096436
 20,9397.56,0.60103,0.73727,1.06316,0.82508,0.84635,0.88487,0.76434,0.50517,0.57093,0.97495,0.00096238,0.00096238,0.00096238
 21,9865.46,0.59967,0.7298,1.05932,0.82435,0.84601,0.88509,0.76508,0.50255,0.56891,0.97282,0.0009604,0.0009604,0.0009604
 22,10328.2,0.59422,0.71917,1.05554,0.82421,0.84651,0.88646,0.7666,0.50108,0.56414,0.97365,0.00095842,0.00095842,0.00095842
 23,10790.8,0.58961,0.71019,1.0554,0.82674,0.8483,0.8881,0.76783,0.49758,0.5603,0.97092,0.00095644,0.00095644,0.00095644
 24,11253.8,0.58536,0.69729,1.05217,0.83952,0.84128,0.8886,0.76988,0.49635,0.55862,0.96996,0.00095446,0.00095446,0.00095446
 25,11716.2,0.58278,0.69802,1.04804,0.84234,0.84168,0.89132,0.771,0.49616,0.55391,0.96949,0.00095248,0.00095248,0.00095248
 26,12178.8,0.57886,0.6897,1.04566,0.85293,0.83682,0.89182,0.77091,0.49339,0.55123,0.96775,0.0009505,0.0009505,0.0009505
 27,12644.5,0.57424,0.67772,1.04268,0.85141,0.83998,0.8931,0.77282,0.49124,0.54865,0.96604,0.00094852,0.00094852,0.00094852
 28,13106.9,0.56989,0.67106,1.0427,0.86569,0.82599,0.89423,0.77372,0.49094,0.54527,0.96495,0.00094654,0.00094654,0.00094654
 29,13569.2,0.56663,0.66424,1.03633,0.87071,0.82732,0.89425,0.77436,0.48888,0.54221,0.96293,0.00094456,0.00094456,0.00094456
 30,14031.5,0.56491,0.65945,1.03474,0.87416,0.82633,0.89702,0.77712,0.4861,0.53997,0.96035,0.00094258,0.00094258,0.00094258
 31,14493.8,0.56111,0.65743,1.03186,0.8724,0.83137,0.8973,0.77804,0.48532,0.53786,0.95917,0.0009406,0.0009406,0.0009406
 32,14955.6,0.55913,0.64898,1.03113,0.87391,0.83375,0.90035,0.77864,0.48281,0.5347,0.95652,0.00093862,0.00093862,0.00093862
 33,15418,0.55545,0.64385,1.03113,0.869,0.83874,0.89926,0.78037,0.48235,0.53183,0.95659,0.00093664,0.00093664,0.00093664
 34,15880.4,0.5508,0.63619,1.03102,0.84745,0.86771,0.90212,0.78125,0.48295,0.52968,0.95805,0.00093466,0.00093466,0.00093466
 35,16342.6,0.54993,0.63723,1.03125,0.85849,0.85634,0.90134,0.78232,0.48162,0.5297,0.95724,0.00093268,0.00093268,0.00093268
 36,16804.7,0.54702,0.62535,1.02967,0.8581,0.85719,0.90146,0.78197,0.4813,0.52816,0.95727,0.0009307,0.0009307,0.0009307
 37,17266.8,0.54243,0.6199,1.02472,0.8606,0.8609,0.90413,0.78324,0.4812,0.52925,0.95703,0.00092872,0.00092872,0.00092872
 38,17729.1,0.54064,0.61272,1.02264,0.86105,0.85965,0.90422,0.78307,0.48115,0.52784,0.9567,0.00092674,0.00092674,0.00092674
 39,18191.2,0.53887,0.60931,1.02123,0.86174,0.85766,0.90456,0.78337,0.48029,0.52615,0.95549,0.00092476,0.00092476,0.00092476
 40,18653.8,0.53139,0.6029,1.01607,0.86435,0.8555,0.90432,0.78401,0.47858,0.52491,0.95361,0.00092278,0.00092278,0.00092278
 41,19116.1,0.53408,0.60101,1.01962,0.86674,0.84936,0.90463,0.78439,0.4777,0.52527,0.95235,0.0009208,0.0009208,0.0009208
 42,19578.3,0.53145,0.59632,1.01591,0.86103,0.85381,0.90495,0.78501,0.47736,0.52465,0.95138,0.00091882,0.00091882,0.00091882
 43,20040.4,0.52868,0.5836,1.01494,0.86249,0.85859,0.90549,0.78575,0.47668,0.5248,0.95091,0.00091684,0.00091684,0.00091684
 44,20504.6,0.52763,0.58484,1.01133,0.86715,0.85773,0.90553,0.78649,0.47562,0.52416,0.94888,0.00091486,0.00091486,0.00091486
 45,20966.7,0.52374,0.57922,1.0097,0.86496,0.85779,0.90613,0.78685,0.4756,0.52425,0.94903,0.00091288,0.00091288,0.00091288
 46,21428.4,0.52154,0.57403,1.01045,0.88001,0.8417,0.90694,0.7888,0.47492,0.5229,0.94851,0.0009109,0.0009109,0.0009109
 47,21890.3,0.52137,0.57439,1.01062,0.88246,0.83694,0.90708,0.79087,0.47572,0.52306,0.94903,0.00090892,0.00090892,0.00090892
 48,22352.5,0.5195,0.56909,1.00541,0.88115,0.83611,0.90723,0.79012,0.47556,0.52117,0.94869,0.00090694,0.00090694,0.00090694
 49,22814.5,0.5188,0.56325,1.00381,0.85185,0.8671,0.90743,0.79159,0.47566,0.51901,0.94893,0.00090496,0.00090496,0.00090496
 50,23276.6,0.51374,0.56129,1.00363,0.88405,0.83195,0.90687,0.79149,0.4756,0.51964,0.94916,0.00090298,0.00090298,0.00090298
 51,23738.5,0.51755,0.55828,1.00527,0.8467,0.86476,0.90715,0.7924,0.47551,0.51703,0.94833,0.000901,0.000901,0.000901
 52,24200.6,0.51481,0.55834,1.00436,0.88885,0.82982,0.90788,0.79203,0.47542,0.51624,0.9479,0.00089902,0.00089902,0.00089902
 53,24662.5,0.5114,0.55254,1.00007,0.88054,0.83439,0.90827,0.79284,0.47445,0.51568,0.94597,0.00089704,0.00089704,0.00089704
 54,25124.8,0.50896,0.54342,0.99716,0.87957,0.8351,0.90828,0.79291,0.47506,0.51621,0.94556,0.00089506,0.00089506,0.00089506
 55,25587.1,0.50462,0.54098,0.99649,0.8772,0.83601,0.90827,0.79414,0.47565,0.51515,0.94563,0.00089308,0.00089308,0.00089308
 56,26049.3,0.50612,0.54491,0.99887,0.88427,0.83355,0.9092,0.79486,0.47587,0.5137,0.94622,0.0008911,0.0008911,0.0008911
 57,26511.4,0.5033,0.53839,0.9971,0.885,0.83479,0.90968,0.79483,0.4749,0.51298,0.94539,0.00088912,0.00088912,0.00088912
 58,26973.4,0.50054,0.53574,0.99372,0.87954,0.83854,0.91048,0.79524,0.47489,0.51276,0.94537,0.00088714,0.00088714,0.00088714
 59,27435.5,0.50108,0.53197,0.99693,0.87807,0.84162,0.91123,0.79554,0.47397,0.51264,0.94511,0.00088516,0.00088516,0.00088516
 60,27897.8,0.50193,0.52879,0.9968,0.87571,0.84718,0.91163,0.79605,0.47396,0.51248,0.94587,0.00088318,0.00088318,0.00088318
 61,28360,0.49687,0.5238,0.99059,0.87557,0.84413,0.91078,0.79437,0.47423,0.51271,0.9461,0.0008812,0.0008812,0.0008812
 62,28822,0.49674,0.51739,0.98821,0.87378,0.84612,0.91117,0.79382,0.47465,0.51305,0.94589,0.00087922,0.00087922,0.00087922
 63,29286.3,0.49831,0.52129,0.99287,0.86326,0.85898,0.91142,0.79623,0.47311,0.51108,0.94409,0.00087724,0.00087724,0.00087724
 64,29748.5,0.49168,0.51482,0.98801,0.86212,0.86008,0.91188,0.79452,0.47257,0.51118,0.94373,0.00087526,0.00087526,0.00087526
 65,30210.8,0.49093,0.51166,0.99064,0.86393,0.85755,0.91178,0.79483,0.47126,0.51308,0.94309,0.00087328,0.00087328,0.00087328
 66,30673.4,0.48996,0.50964,0.98871,0.86439,0.85836,0.9118,0.79626,0.47112,0.51114,0.94228,0.0008713,0.0008713,0.0008713
 67,31135.8,0.48725,0.50594,0.98582,0.86643,0.85344,0.91203,0.79811,0.4705,0.50867,0.94157,0.00086932,0.00086932,0.00086932
 68,31597.9,0.48673,0.5033,0.988,0.89025,0.83037,0.9119,0.79859,0.47006,0.50904,0.94153,0.00086734,0.00086734,0.00086734
 69,32060.2,0.48656,0.50139,0.98504,0.88768,0.83192,0.91147,0.79808,0.4689,0.50962,0.94109,0.00086536,0.00086536,0.00086536
 70,32522.4,0.48348,0.50085,0.98434,0.89171,0.82874,0.9115,0.79799,0.46792,0.5102,0.94141,0.00086338,0.00086338,0.00086338
 71,32984.4,0.4847,0.49669,0.98437,0.886,0.82957,0.91131,0.79849,0.46777,0.51222,0.94197,0.0008614,0.0008614,0.0008614
 72,33446.5,0.48186,0.49734,0.98176,0.8758,0.84131,0.91079,0.79711,0.46655,0.51396,0.941,0.00085942,0.00085942,0.00085942
 73,33908.6,0.48037,0.49197,0.98147,0.87872,0.84683,0.91038,0.79767,0.46532,0.5139,0.94027,0.00085744,0.00085744,0.00085744
 74,34370.7,0.48222,0.49279,0.9825,0.87941,0.8471,0.91026,0.79635,0.46446,0.51482,0.93941,0.00085546,0.00085546,0.00085546
 75,34833,0.47775,0.48748,0.97907,0.87694,0.84631,0.91041,0.79614,0.46376,0.51531,0.93964,0.00085348,0.00085348,0.00085348
 76,35295.1,0.47712,0.48701,0.97857,0.87412,0.84681,0.91055,0.79561,0.46292,0.51465,0.93928,0.0008515,0.0008515,0.0008515
 77,35757.2,0.47445,0.47909,0.97752,0.87748,0.83806,0.91043,0.79662,0.46253,0.51421,0.93939,0.00084952,0.00084952,0.00084952
 78,36219.4,0.47499,0.47603,0.97722,0.8833,0.83726,0.90997,0.79638,0.46193,0.51479,0.93871,0.00084754,0.00084754,0.00084754
 79,36681.7,0.47372,0.47831,0.9738,0.87848,0.83876,0.90969,0.79656,0.46402,0.5171,0.93979,0.00084556,0.00084556,0.00084556
 80,37146.3,0.47188,0.4755,0.97527,0.88542,0.83955,0.90996,0.7963,0.46453,0.52173,0.94004,0.00084358,0.00084358,0.00084358
 81,37608.1,0.47261,0.47433,0.97704,0.88365,0.83769,0.90964,0.79562,0.46532,0.52287,0.94027,0.0008416,0.0008416,0.0008416
 82,38069.9,0.47042,0.47008,0.97466,0.88444,0.83781,0.91002,0.79564,0.46497,0.52616,0.94048,0.00083962,0.00083962,0.00083962
 83,38531.6,0.46762,0.46805,0.97435,0.88983,0.83006,0.90899,0.7961,0.46381,0.52777,0.94002,0.00083764,0.00083764,0.00083764
 84,38993.3,0.4676,0.46594,0.97384,0.88759,0.82934,0.90852,0.79404,0.46423,0.52943,0.94102,0.00083566,0.00083566,0.00083566
 85,39455,0.46796,0.46293,0.97068,0.8893,0.83036,0.90833,0.79482,0.46383,0.53027,0.94096,0.00083368,0.00083368,0.00083368
 86,39916.8,0.46728,0.46179,0.9702,0.88996,0.83023,0.90837,0.79495,0.46383,0.53168,0.9396,0.0008317,0.0008317,0.0008317
 87,40378.8,0.46298,0.46029,0.9684,0.88906,0.82899,0.90866,0.795,0.46383,0.53253,0.93945,0.00082972,0.00082972,0.00082972
 88,40840.9,0.46303,0.45772,0.96827,0.88776,0.82946,0.90865,0.7949,0.46352,0.53185,0.93884,0.00082774,0.00082774,0.00082774
 89,41303.3,0.46176,0.45664,0.97013,0.88792,0.82978,0.90767,0.79487,0.46435,0.53372,0.93888,0.00082576,0.00082576,0.00082576
 90,41765.4,0.46083,0.45357,0.97121,0.89043,0.82816,0.90783,0.79412,0.46422,0.53632,0.93912,0.00082378,0.00082378,0.00082378
 91,42228,0.46107,0.45139,0.96927,0.86802,0.84779,0.90782,0.79372,0.46463,0.539,0.93956,0.0008218,0.0008218,0.0008218
 92,42690,0.45929,0.45307,0.96682,0.86522,0.85161,0.90737,0.79374,0.46378,0.5376,0.93943,0.00081982,0.00081982,0.00081982
 93,43152,0.46016,0.45065,0.96842,0.87161,0.84493,0.90652,0.79386,0.46465,0.5382,0.94029,0.00081784,0.00081784,0.00081784
 94,43613.9,0.45718,0.44908,0.96704,0.86077,0.85589,0.90606,0.79442,0.46348,0.53736,0.94013,0.00081586,0.00081586,0.00081586
 95,44076,0.45887,0.44694,0.96849,0.86092,0.85986,0.90599,0.79477,0.46362,0.53655,0.93988,0.00081388,0.00081388,0.00081388
 96,44538.2,0.45504,0.44399,0.96595,0.8591,0.85556,0.90346,0.793,0.46296,0.53809,0.93997,0.0008119,0.0008119,0.0008119
 97,45000.2,0.45516,0.44315,0.96648,0.85801,0.85308,0.90244,0.79305,0.46338,0.53927,0.94173,0.00080992,0.00080992,0.00080992
 98,45462.4,0.45233,0.43965,0.96518,0.86607,0.84775,0.90176,0.79221,0.46451,0.53978,0.94244,0.00080794,0.00080794,0.00080794
 99,45927,0.45221,0.43739,0.96207,0.86603,0.85345,0.90164,0.79354,0.46665,0.54085,0.94437,0.00080596,0.00080596,0.00080596
 100,46389,0.44955,0.43427,0.96153,0.86753,0.8577,0.90403,0.79507,0.46633,0.53894,0.94464,0.00080398,0.00080398,0.00080398
 101,46851,0.44909,0.43242,0.96049,0.87025,0.8561,0.90428,0.79425,0.46643,0.5384,0.94503,0.000802,0.000802,0.000802
 102,47313.1,0.44779,0.42907,0.95638,0.87006,0.86017,0.9048,0.79615,0.46807,0.54049,0.94549,0.00080002,0.00080002,0.00080002
 103,47775.2,0.45088,0.43586,0.96374,0.86912,0.86268,0.90498,0.79514,0.46953,0.54258,0.94677,0.00079804,0.00079804,0.00079804
 104,48237.2,0.4486,0.43123,0.95881,0.88179,0.84919,0.90513,0.79491,0.47222,0.54326,0.94855,0.00079606,0.00079606,0.00079606
 105,48699.2,0.44872,0.42823,0.95905,0.87909,0.84653,0.90294,0.79446,0.47171,0.54543,0.9477,0.00079408,0.00079408,0.00079408
 106,49161.5,0.44819,0.42853,0.96006,0.88599,0.84255,0.90283,0.79375,0.47042,0.54522,0.94576,0.0007921,0.0007921,0.0007921
 107,49623.6,0.44441,0.426,0.95773,0.87972,0.84742,0.90296,0.79553,0.4701,0.54547,0.94559,0.00079012,0.00079012,0.00079012
 108,50085.9,0.44308,0.42388,0.95475,0.8702,0.85277,0.90206,0.79537,0.46946,0.54852,0.94544,0.00078814,0.00078814,0.00078814
 109,50547.5,0.44443,0.4185,0.95548,0.8828,0.83732,0.90172,0.7944,0.46905,0.5499,0.94529,0.00078616,0.00078616,0.00078616
 110,51010.2,0.44387,0.42138,0.95435,0.86257,0.86004,0.90129,0.79357,0.46927,0.55072,0.94507,0.00078418,0.00078418,0.00078418
 111,51472.4,0.44359,0.42036,0.95318,0.88122,0.84597,0.90106,0.79301,0.46768,0.55016,0.9435,0.0007822,0.0007822,0.0007822
 112,51934.3,0.44207,0.41848,0.95537,0.88093,0.84526,0.90148,0.7944,0.46795,0.54995,0.94429,0.00078022,0.00078022,0.00078022
 113,52397.2,0.44164,0.41628,0.95591,0.87769,0.84522,0.90114,0.79354,0.46791,0.55027,0.9454,0.00077824,0.00077824,0.00077824
 114,52859.2,0.44109,0.41413,0.95581,0.8781,0.84457,0.90005,0.79341,0.4674,0.55128,0.94613,0.00077626,0.00077626,0.00077626
 115,53321.7,0.4415,0.41856,0.95716,0.88507,0.83603,0.90013,0.79336,0.46945,0.55186,0.94752,0.00077428,0.00077428,0.00077428
 116,53783.6,0.43807,0.41112,0.9522,0.89038,0.83195,0.89959,0.7949,0.4693,0.55269,0.94728,0.0007723,0.0007723,0.0007723
 117,54247.7,0.4357,0.41175,0.95447,0.88819,0.83169,0.89862,0.7935,0.46842,0.55237,0.94603,0.00077032,0.00077032,0.00077032
 118,54709.8,0.43707,0.41214,0.95187,0.87947,0.84319,0.89929,0.79291,0.46871,0.5529,0.94508,0.00076834,0.00076834,0.00076834
--- a/fashion_yolo/fashionpeida_exp.ipynb
+++ b/fashion_yolo/fashionpeida_exp.ipynb
@ -567,28 +567,28 @@
     "output_type": "stream",
     "text": [
      "Ultralytics 8.3.225 🚀 Python-3.10.18 torch-2.9.1+cu128 CUDA:0 (NVIDIA GeForce RTX 5090, 32109MiB)\n",
-      "Model summary (fused): 92 layers, 25,849,024 parameters, 0 gradients, 78.7 GFLOPs\n",
+      "Model summary (fused): 92 layers, 25,842,076 parameters, 0 gradients, 78.7 GFLOPs\n",
      "\n",
-      "\u001b[34m\u001b[1mPyTorch:\u001b[0m starting from '/home/cuuva/experiment/fashion_yolo/fashionpedia_exp/yolov8m_fashion_final2/weights/best_fashion_16class.pt' with input shape (1, 3, 640, 640) BCHW and output shape(s) (1, 20, 8400) (49.6 MB)\n",
+      "\u001b[34m\u001b[1mPyTorch:\u001b[0m starting from '/home/cuuva/git/Detection_Experiment/fashion_yolo/finetuning_exp/finetuned_exp/yolov8m_finetuned_4class/weights/best_finetuned_clothes.pt' with input shape (1, 3, 640, 640) BCHW and output shape(s) (1, 8, 8400) (49.6 MB)\n",
      "\n",
      "\u001b[34m\u001b[1mONNX:\u001b[0m starting export with onnx 1.19.1 opset 20...\n",
      "\u001b[34m\u001b[1mONNX:\u001b[0m slimming with onnxslim 0.1.71...\n",
-      "\u001b[34m\u001b[1mONNX:\u001b[0m export success ✅ 0.8s, saved as '/home/cuuva/experiment/fashion_yolo/fashionpedia_exp/yolov8m_fashion_final2/weights/best_fashion_16class.onnx' (98.9 MB)\n",
+      "\u001b[34m\u001b[1mONNX:\u001b[0m export success ✅ 0.6s, saved as '/home/cuuva/git/Detection_Experiment/fashion_yolo/finetuning_exp/finetuned_exp/yolov8m_finetuned_4class/weights/best_finetuned_clothes.onnx' (98.8 MB)\n",
      "\n",
-      "Export complete (1.1s)\n",
+      "Export complete (0.7s)\n",
-      "Results saved to \u001b[1m/home/cuuva/experiment/fashion_yolo/fashionpedia_exp/yolov8m_fashion_final2/weights\u001b[0m\n",
+      "Results saved to \u001b[1m/home/cuuva/git/Detection_Experiment/fashion_yolo/finetuning_exp/finetuned_exp/yolov8m_finetuned_4class/weights\u001b[0m\n",
-      "Predict:         yolo predict task=detect model=/home/cuuva/experiment/fashion_yolo/fashionpedia_exp/yolov8m_fashion_final2/weights/best_fashion_16class.onnx imgsz=640  \n",
+      "Predict:         yolo predict task=detect model=/home/cuuva/git/Detection_Experiment/fashion_yolo/finetuning_exp/finetuned_exp/yolov8m_finetuned_4class/weights/best_finetuned_clothes.onnx imgsz=640  \n",
-      "Validate:        yolo val task=detect model=/home/cuuva/experiment/fashion_yolo/fashionpedia_exp/yolov8m_fashion_final2/weights/best_fashion_16class.onnx imgsz=640 data=/home/cuuva/experiment/datasets/fashionpedia_yolo/fashionpedia_reduced.yaml  \n",
+      "Validate:        yolo val task=detect model=/home/cuuva/git/Detection_Experiment/fashion_yolo/finetuning_exp/finetuned_exp/yolov8m_finetuned_4class/weights/best_finetuned_clothes.onnx imgsz=640 data=/home/cuuva/git/Detection_Experiment/fashion_yolo/finetuning_exp/finetuning_dataset.yaml  \n",
      "Visualize:       https://netron.app\n"
     ]
    },
    {
     "data": {
      "text/plain": [
-       "'/home/cuuva/experiment/fashion_yolo/fashionpedia_exp/yolov8m_fashion_final2/weights/best_fashion_16class.onnx'"
+       "'/home/cuuva/git/Detection_Experiment/fashion_yolo/finetuning_exp/finetuned_exp/yolov8m_finetuned_4class/weights/best_finetuned_clothes.onnx'"
      ]
     },
-     "execution_count": 1,
+     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
--- a/fashion_yolo/finetuning_exp/finetuned_exp/yolov8m_finetuned_16class/args.yaml
+++ b/fashion_yolo/finetuning_exp/finetuned_exp/yolov8m_finetuned_16class/args.yaml
@ -0,0 +1,106 @@
 task: detect
 mode: train
 model: yolov8m.pt
 data: /home/cuuva/git/Detection_Experiment/fashion_yolo/finetuning_exp/finetuning_dataset.yaml
 epochs: 200
 time: null
 patience: 50
 batch: -1
 imgsz: 640
 save: true
 save_period: -1
 cache: false
 device: '0'
 workers: 8
 project: finetuned_exp
 name: yolov8m_finetuned_16class
 exist_ok: false
 pretrained: true
 optimizer: AdamW
 verbose: true
 seed: 0
 deterministic: true
 single_cls: false
 rect: false
 cos_lr: false
 close_mosaic: 10
 resume: false
 amp: true
 fraction: 1.0
 profile: false
 freeze: null
 multi_scale: false
 compile: false
 overlap_mask: true
 mask_ratio: 4
 dropout: 0.0
 val: true
 split: val
 save_json: false
 conf: null
 iou: 0.7
 max_det: 300
 half: false
 dnn: false
 plots: true
 source: null
 vid_stride: 1
 stream_buffer: false
 visualize: false
 augment: false
 agnostic_nms: false
 classes: null
 retina_masks: false
 embed: null
 show: false
 save_frames: false
 save_txt: false
 save_conf: false
 save_crop: false
 show_labels: true
 show_conf: true
 show_boxes: true
 line_width: null
 format: torchscript
 keras: false
 optimize: false
 int8: false
 dynamic: false
 simplify: true
 opset: null
 workspace: null
 nms: false
 lr0: 0.001
 lrf: 0.01
 momentum: 0.937
 weight_decay: 0.0005
 warmup_epochs: 3.0
 warmup_momentum: 0.8
 warmup_bias_lr: 0.1
 box: 7.5
 cls: 0.5
 dfl: 1.5
 pose: 12.0
 kobj: 1.0
 nbs: 64
 hsv_h: 0.015
 hsv_s: 0.7
 hsv_v: 0.4
 degrees: 0.0
 translate: 0.1
 scale: 0.5
 shear: 0.0
 perspective: 0.0
 flipud: 0.0
 fliplr: 0.5
 bgr: 0.0
 mosaic: 1.0
 mixup: 0.0
 cutmix: 0.0
 copy_paste: 0.0
 copy_paste_mode: flip
 auto_augment: randaugment
 erasing: 0.4
 cfg: null
 tracker: botsort.yaml
 save_dir: /home/cuuva/git/Detection_Experiment/fashion_yolo/finetuning_exp/finetuned_exp/yolov8m_finetuned_16class
--- a/fashion_yolo/finetuning_exp/finetuned_exp/yolov8m_finetuned_16class/results.csv
+++ b/fashion_yolo/finetuning_exp/finetuned_exp/yolov8m_finetuned_16class/results.csv
@ -0,0 +1,64 @@
 epoch,time,train/box_loss,train/cls_loss,train/dfl_loss,metrics/precision(B),metrics/recall(B),metrics/mAP50(B),metrics/mAP50-95(B),val/box_loss,val/cls_loss,val/dfl_loss,lr/pg0,lr/pg1,lr/pg2
 1,3.78148,1.33469,3.80034,1.63059,0.73796,0.29947,0.37698,0.31586,0.52953,2.60886,1.0286,0.08911,0.00011,0.00011
 2,7.56087,0.60945,0.99123,0.97196,0.68549,0.36595,0.43755,0.32073,0.70291,3.06519,1.10377,0.0772289,0.000228861,0.000228861
 3,11.4202,0.51229,0.67713,0.90343,0.71504,0.3875,0.44526,0.29532,0.6803,10.726,1.0683,0.0653465,0.000346535,0.000346535
 4,15.1807,0.52074,0.63101,0.90427,0.65366,0.60284,0.48227,0.29742,0.60168,3.51049,1.04707,0.053463,0.00046302,0.00046302
 5,18.9709,0.50722,0.61079,0.90078,0.73198,0.57424,0.49323,0.44174,0.39059,1.28564,0.86867,0.0415783,0.000578318,0.000578318
 6,22.8854,0.49552,0.60369,0.89034,0.66355,0.59467,0.52354,0.41981,0.42398,1.57978,0.89592,0.0296924,0.000692427,0.000692427
 7,26.6285,0.46947,0.55428,0.87874,0.61267,0.6248,0.43648,0.35161,0.44943,1.45805,0.92086,0.0178053,0.000805349,0.000805349
 8,30.3691,0.45534,0.52409,0.88231,0.6246,0.63572,0.45639,0.33998,0.44835,1.93294,0.95281,0.00591708,0.000917082,0.000917082
 9,34.1485,0.48601,0.56182,0.89716,0.59408,0.61932,0.4629,0.32455,0.45378,1.53813,0.95085,0.0009604,0.0009604,0.0009604
 10,37.921,0.48816,0.5445,0.89966,0.63695,0.63235,0.49343,0.39283,0.48469,3.00655,0.96584,0.00095545,0.00095545,0.00095545
 11,41.6616,0.47282,0.52913,0.884,0.70406,0.69102,0.48837,0.37312,0.45323,2.1499,0.94868,0.0009505,0.0009505,0.0009505
 12,45.4309,0.47726,0.51821,0.88777,0.67332,0.63697,0.55051,0.43409,0.42691,1.43266,0.8991,0.00094555,0.00094555,0.00094555
 13,49.3216,0.47322,0.50814,0.87982,0.69371,0.70732,0.5492,0.50737,0.39498,1.77262,0.92069,0.0009406,0.0009406,0.0009406
 14,53.1589,0.45547,0.48886,0.87557,0.65904,0.66094,0.52163,0.47477,0.37669,1.23177,0.87385,0.00093565,0.00093565,0.00093565
 15,56.9039,0.44566,0.48189,0.87569,0.60504,0.54129,0.43602,0.38708,0.41243,1.03501,0.89825,0.0009307,0.0009307,0.0009307
 16,60.6523,0.44933,0.502,0.88999,0.37379,0.62594,0.43582,0.34925,0.42962,1.78191,0.91027,0.00092575,0.00092575,0.00092575
 17,64.4368,0.43722,0.48608,0.8777,0.63457,0.64845,0.4299,0.32622,0.37782,1.73409,0.86714,0.0009208,0.0009208,0.0009208
 18,68.1841,0.42233,0.46144,0.8685,0.69631,0.66645,0.49723,0.42279,0.37372,1.10271,0.87895,0.00091585,0.00091585,0.00091585
 19,71.9377,0.40829,0.4449,0.86003,0.66101,0.60477,0.48363,0.40767,0.39504,1.1244,0.89122,0.0009109,0.0009109,0.0009109
 20,75.6855,0.409,0.45819,0.8668,0.70192,0.61673,0.54394,0.46983,0.39188,0.93353,0.88193,0.00090595,0.00090595,0.00090595
 21,79.5951,0.40671,0.44567,0.8634,0.74314,0.50871,0.54476,0.47298,0.44147,0.96029,0.90582,0.000901,0.000901,0.000901
 22,83.361,0.41313,0.45403,0.87557,0.4601,0.47457,0.45017,0.33278,0.53013,1.32302,0.96023,0.00089605,0.00089605,0.00089605
 23,87.1423,0.41917,0.45142,0.86546,0.38438,0.60481,0.46156,0.42506,0.37881,0.90096,0.87146,0.0008911,0.0008911,0.0008911
 24,90.9445,0.4293,0.4509,0.8811,0.72349,0.55997,0.49705,0.3744,0.42867,1.08573,0.91291,0.00088615,0.00088615,0.00088615
 25,94.7317,0.41275,0.45251,0.86412,0.6813,0.62841,0.46227,0.37507,0.38894,1.02401,0.89599,0.0008812,0.0008812,0.0008812
 26,98.5067,0.40656,0.43857,0.86411,0.69541,0.62841,0.4795,0.409,0.38527,1.14159,0.87603,0.00087625,0.00087625,0.00087625
 27,102.282,0.4068,0.44555,0.86759,0.71762,0.55763,0.48642,0.39518,0.35101,1.24645,0.86887,0.0008713,0.0008713,0.0008713
 28,106.172,0.39003,0.42322,0.85955,0.46189,0.69735,0.50023,0.34984,0.42298,1.1147,0.91119,0.00086635,0.00086635,0.00086635
 29,109.981,0.39569,0.44981,0.85652,0.76322,0.64827,0.55903,0.40588,0.37011,0.86242,0.88865,0.0008614,0.0008614,0.0008614
 30,113.742,0.39019,0.41622,0.8555,0.4939,0.58227,0.50257,0.37889,0.37393,1.15246,0.89575,0.00085645,0.00085645,0.00085645
 31,117.525,0.37475,0.41644,0.84947,0.39663,0.59492,0.4467,0.33685,0.40234,1.25879,0.93017,0.0008515,0.0008515,0.0008515
 32,121.325,0.37289,0.40073,0.84581,0.69496,0.62941,0.4938,0.36733,0.39513,1.27688,0.93298,0.00084655,0.00084655,0.00084655
 33,125.089,0.3722,0.42107,0.85543,0.75471,0.64346,0.52761,0.39675,0.36185,1.056,0.90608,0.0008416,0.0008416,0.0008416
 34,128.849,0.37924,0.41992,0.84365,0.7313,0.66103,0.52418,0.42032,0.31768,0.9188,0.87799,0.00083665,0.00083665,0.00083665
 35,132.714,0.36662,0.38708,0.83892,0.73818,0.52841,0.47963,0.39382,0.33542,1.05785,0.86738,0.0008317,0.0008317,0.0008317
 36,136.535,0.37632,0.38432,0.8546,0.66502,0.59938,0.51885,0.40912,0.34029,0.97035,0.87475,0.00082675,0.00082675,0.00082675
 37,140.338,0.37271,0.38587,0.84571,0.5174,0.63531,0.52793,0.43814,0.36554,1.07988,0.88498,0.0008218,0.0008218,0.0008218
 38,144.134,0.36901,0.40024,0.85047,0.71707,0.59129,0.56226,0.42374,0.40125,0.99763,0.9244,0.00081685,0.00081685,0.00081685
 39,147.934,0.39434,0.43089,0.8593,0.55562,0.55194,0.40805,0.34409,0.44815,1.09204,0.95664,0.0008119,0.0008119,0.0008119
 40,151.698,0.37155,0.38967,0.84658,0.68503,0.54394,0.52193,0.41761,0.47835,1.08346,0.97604,0.00080695,0.00080695,0.00080695
 41,155.498,0.36644,0.38629,0.84955,0.67577,0.54583,0.50672,0.42354,0.40004,0.98989,0.93725,0.000802,0.000802,0.000802
 42,159.415,0.35604,0.38401,0.84641,0.60819,0.56554,0.41907,0.34978,0.37567,0.99076,0.91391,0.00079705,0.00079705,0.00079705
 43,163.217,0.35579,0.37579,0.85306,0.61134,0.62386,0.47743,0.37828,0.45271,1.08443,0.93421,0.0007921,0.0007921,0.0007921
 44,166.984,0.35405,0.37357,0.83472,0.75872,0.4296,0.47769,0.34909,0.46733,1.45015,0.94675,0.00078715,0.00078715,0.00078715
 45,170.756,0.3442,0.36419,0.84202,0.65868,0.56114,0.45587,0.37317,0.37198,1.31101,0.89396,0.0007822,0.0007822,0.0007822
 46,174.56,0.35409,0.35863,0.83752,0.61909,0.56445,0.41229,0.35977,0.36592,1.14361,0.86921,0.00077725,0.00077725,0.00077725
 47,178.339,0.35101,0.37277,0.83852,0.62703,0.55114,0.43938,0.38073,0.3484,1.08414,0.85939,0.0007723,0.0007723,0.0007723
 48,182.107,0.36916,0.36519,0.84566,0.71307,0.48674,0.46579,0.38478,0.36118,1.03343,0.8817,0.00076735,0.00076735,0.00076735
 49,185.88,0.34978,0.35519,0.83506,0.71914,0.36925,0.43888,0.34882,0.37716,1.13054,0.90382,0.0007624,0.0007624,0.0007624
 50,189.752,0.36345,0.3697,0.8442,0.69832,0.43709,0.44508,0.36078,0.37054,1.12954,0.89464,0.00075745,0.00075745,0.00075745
 51,193.558,0.35846,0.36267,0.84488,0.6433,0.60038,0.46064,0.34586,0.39035,1.20544,0.9171,0.0007525,0.0007525,0.0007525
 52,197.33,0.35449,0.37803,0.84451,0.63339,0.58406,0.44893,0.33556,0.40868,1.02716,0.92945,0.00074755,0.00074755,0.00074755
 53,201.139,0.35648,0.36618,0.84817,0.63325,0.55909,0.44089,0.3366,0.42693,1.2246,0.9431,0.0007426,0.0007426,0.0007426
 54,204.91,0.35324,0.36185,0.84395,0.42795,0.53118,0.44295,0.32709,0.39309,1.33043,0.92094,0.00073765,0.00073765,0.00073765
 55,208.677,0.34405,0.35182,0.84253,0.7537,0.48704,0.51626,0.39089,0.39022,0.98264,0.91645,0.0007327,0.0007327,0.0007327
 56,212.481,0.32941,0.34251,0.83695,0.61177,0.52932,0.45094,0.34772,0.37838,1.08123,0.88894,0.00072775,0.00072775,0.00072775
 57,216.394,0.3348,0.33981,0.83636,0.71788,0.45155,0.4911,0.40562,0.33227,0.97807,0.86522,0.0007228,0.0007228,0.0007228
 58,220.183,0.34344,0.34474,0.84528,0.61592,0.58481,0.44577,0.38492,0.32137,1.06028,0.86325,0.00071785,0.00071785,0.00071785
 59,223.99,0.33839,0.34634,0.8355,0.62415,0.53198,0.43579,0.3386,0.36025,1.21482,0.88833,0.0007129,0.0007129,0.0007129
 60,227.795,0.33307,0.33585,0.83415,0.72749,0.50856,0.47991,0.37027,0.37971,1.17272,0.91263,0.00070795,0.00070795,0.00070795
 61,231.601,0.32663,0.32414,0.83285,0.69812,0.5339,0.41448,0.32622,0.38985,1.3369,0.91243,0.000703,0.000703,0.000703
 62,235.371,0.33762,0.33916,0.83625,0.69725,0.47572,0.46259,0.36795,0.34215,1.2375,0.88089,0.00069805,0.00069805,0.00069805
 63,239.185,0.33312,0.33786,0.83502,0.67084,0.49598,0.49992,0.43129,0.34499,1.09861,0.8657,0.0006931,0.0006931,0.0006931
--- a/fashion_yolo/finetuning_exp/finetuned_exp/yolov8m_finetuned_16class2/args.yaml
+++ b/fashion_yolo/finetuning_exp/finetuned_exp/yolov8m_finetuned_16class2/args.yaml
@ -0,0 +1,106 @@
 task: detect
 mode: train
 model: /home/cuuva/git/Detection_Experiment/fashion_yolo/fashionpedia_exp/yolov8m_fashion_final/weights/best_fashion_16class.pt
 data: /home/cuuva/git/Detection_Experiment/fashion_yolo/finetuning_exp/finetuning_dataset.yaml
 epochs: 200
 time: null
 patience: 50
 batch: -1
 imgsz: 640
 save: true
 save_period: -1
 cache: false
 device: '0'
 workers: 8
 project: finetuned_exp
 name: yolov8m_finetuned_16class2
 exist_ok: false
 pretrained: true
 optimizer: AdamW
 verbose: true
 seed: 0
 deterministic: true
 single_cls: false
 rect: false
 cos_lr: false
 close_mosaic: 10
 resume: false
 amp: true
 fraction: 1.0
 profile: false
 freeze: null
 multi_scale: false
 compile: false
 overlap_mask: true
 mask_ratio: 4
 dropout: 0.0
 val: true
 split: val
 save_json: false
 conf: null
 iou: 0.7
 max_det: 300
 half: false
 dnn: false
 plots: true
 source: null
 vid_stride: 1
 stream_buffer: false
 visualize: false
 augment: false
 agnostic_nms: false
 classes: null
 retina_masks: false
 embed: null
 show: false
 save_frames: false
 save_txt: false
 save_conf: false
 save_crop: false
 show_labels: true
 show_conf: true
 show_boxes: true
 line_width: null
 format: torchscript
 keras: false
 optimize: false
 int8: false
 dynamic: false
 simplify: true
 opset: null
 workspace: null
 nms: false
 lr0: 0.001
 lrf: 0.01
 momentum: 0.937
 weight_decay: 0.0005
 warmup_epochs: 3.0
 warmup_momentum: 0.8
 warmup_bias_lr: 0.1
 box: 7.5
 cls: 0.5
 dfl: 1.5
 pose: 12.0
 kobj: 1.0
 nbs: 64
 hsv_h: 0.015
 hsv_s: 0.7
 hsv_v: 0.4
 degrees: 0.0
 translate: 0.1
 scale: 0.5
 shear: 0.0
 perspective: 0.0
 flipud: 0.0
 fliplr: 0.5
 bgr: 0.0
 mosaic: 1.0
 mixup: 0.0
 cutmix: 0.0
 copy_paste: 0.0
 copy_paste_mode: flip
 auto_augment: randaugment
 erasing: 0.4
 cfg: null
 tracker: botsort.yaml
 save_dir: /home/cuuva/git/Detection_Experiment/fashion_yolo/finetuning_exp/finetuned_exp/yolov8m_finetuned_16class2
--- a/fashion_yolo/finetuning_exp/finetuned_exp/yolov8m_finetuned_16class2/results.csv
+++ b/fashion_yolo/finetuning_exp/finetuned_exp/yolov8m_finetuned_16class2/results.csv
@ -0,0 +1,52 @@
 epoch,time,train/box_loss,train/cls_loss,train/dfl_loss,metrics/precision(B),metrics/recall(B),metrics/mAP50(B),metrics/mAP50-95(B),val/box_loss,val/cls_loss,val/dfl_loss,lr/pg0,lr/pg1,lr/pg2
 1,3.77833,0.79471,2.21157,1.07614,0.82283,0.67952,0.65168,0.58361,0.2879,0.81676,0.80504,0.09307,7e-05,7e-05
 2,7.58108,0.48038,0.64049,0.89175,0.23706,0.15,0.18257,0.16018,1.40033,5.81627,1.69739,0.0851493,0.000149257,0.000149257
 3,11.3715,0.44208,0.54362,0.8741,0.44318,0.15,0.21877,0.18264,1.00022,6.42565,1.30708,0.0772277,0.000227723,0.000227723
 4,15.1751,0.43437,0.50955,0.86559,0.60381,0.22004,0.44609,0.41531,0.40947,3.81086,0.99871,0.0693054,0.000305396,0.000305396
 5,19.007,0.4183,0.48185,0.85746,0.48998,0.45297,0.52371,0.45405,0.30928,1.71318,0.88152,0.0613823,0.000382278,0.000382278
 6,22.831,0.39603,0.45766,0.85549,0.4776,0.52709,0.51232,0.41371,0.31855,1.30563,0.87961,0.0534584,0.000458368,0.000458368
 7,26.6293,0.39399,0.44335,0.85925,0.71274,0.56236,0.51689,0.43259,0.30069,1.20965,0.85695,0.0455337,0.000533665,0.000533665
 8,30.4355,0.37853,0.42514,0.85528,0.73366,0.69107,0.53786,0.48311,0.27062,0.88329,0.83511,0.0376082,0.00060817,0.00060817
 9,34.2788,0.3629,0.41388,0.84647,0.75546,0.59562,0.55273,0.52939,0.27376,0.98462,0.8318,0.0296819,0.000681884,0.000681884
 10,38.0895,0.36652,0.40245,0.85834,0.73261,0.7107,0.53099,0.51115,0.25461,0.82192,0.81761,0.0217548,0.000754805,0.000754805
 11,41.8874,0.37154,0.39447,0.84666,0.76387,0.72008,0.57501,0.56018,0.23619,0.80714,0.81105,0.0138269,0.000826935,0.000826935
 12,45.7351,0.35457,0.38923,0.8447,0.75858,0.68199,0.55223,0.52836,0.24854,0.83378,0.81443,0.00589827,0.000898272,0.000898272
 13,49.5798,0.35328,0.37496,0.84146,0.77044,0.60038,0.59242,0.55918,0.22654,0.82054,0.81329,0.0009406,0.0009406,0.0009406
 14,53.3884,0.34215,0.36192,0.84175,0.77595,0.61781,0.57172,0.50778,0.24315,0.84383,0.81304,0.00093565,0.00093565,0.00093565
 15,57.2074,0.34514,0.36506,0.84374,0.7598,0.59745,0.57575,0.51644,0.23654,0.81983,0.7962,0.0009307,0.0009307,0.0009307
 16,61.0579,0.34922,0.36886,0.84357,0.72997,0.57645,0.58202,0.52647,0.24789,0.86689,0.78555,0.00092575,0.00092575,0.00092575
 17,64.904,0.33659,0.35582,0.84074,0.74304,0.5353,0.53709,0.4965,0.23814,0.89671,0.78294,0.0009208,0.0009208,0.0009208
 18,68.7503,0.33479,0.35639,0.83934,0.73251,0.54903,0.52264,0.49088,0.25039,0.88331,0.78314,0.00091585,0.00091585,0.00091585
 19,72.5669,0.32219,0.34158,0.83553,0.76031,0.57904,0.54172,0.5014,0.23453,0.87966,0.7813,0.0009109,0.0009109,0.0009109
 20,76.386,0.31115,0.33391,0.82998,0.82936,0.58105,0.59831,0.54499,0.23444,0.88116,0.78877,0.00090595,0.00090595,0.00090595
 21,80.2292,0.30777,0.32447,0.82886,0.77372,0.63129,0.60019,0.55856,0.2424,0.95364,0.79713,0.000901,0.000901,0.000901
 22,84.0425,0.31715,0.32831,0.82225,0.77087,0.64659,0.59484,0.55514,0.23616,0.93773,0.79265,0.00089605,0.00089605,0.00089605
 23,87.872,0.32344,0.33185,0.83772,0.76405,0.62121,0.58662,0.54822,0.22635,0.86867,0.78582,0.0008911,0.0008911,0.0008911
 24,91.6934,0.29904,0.30918,0.82789,0.75365,0.65785,0.56653,0.54192,0.2256,1.02833,0.78277,0.00088615,0.00088615,0.00088615
 25,95.5542,0.30904,0.3227,0.83325,0.72991,0.58674,0.52557,0.49026,0.23207,1.07519,0.77621,0.0008812,0.0008812,0.0008812
 26,99.4246,0.31439,0.32609,0.83633,0.67262,0.65758,0.51495,0.48505,0.23068,1.05546,0.77321,0.00087625,0.00087625,0.00087625
 27,103.252,0.32393,0.31774,0.83351,0.72833,0.63511,0.54074,0.50772,0.2415,1.11905,0.77226,0.0008713,0.0008713,0.0008713
 28,107.072,0.3126,0.30388,0.82889,0.74399,0.62573,0.57397,0.54305,0.23596,1.01647,0.79593,0.00086635,0.00086635,0.00086635
 29,110.932,0.30688,0.32481,0.82411,0.7982,0.57745,0.562,0.51721,0.2412,0.8918,0.80653,0.0008614,0.0008614,0.0008614
 30,114.748,0.31745,0.30503,0.82192,0.7413,0.59721,0.5606,0.52255,0.24307,1.00615,0.81755,0.00085645,0.00085645,0.00085645
 31,118.605,0.30967,0.30679,0.83172,0.74443,0.58319,0.51612,0.4751,0.24831,1.12382,0.82134,0.0008515,0.0008515,0.0008515
 32,122.437,0.29943,0.29849,0.82889,0.72564,0.60961,0.52207,0.48467,0.24063,1.0824,0.82329,0.00084655,0.00084655,0.00084655
 33,126.296,0.29808,0.29118,0.82238,0.68827,0.62049,0.54917,0.49879,0.25117,1.16063,0.83648,0.0008416,0.0008416,0.0008416
 34,130.158,0.30958,0.30369,0.82733,0.74312,0.575,0.52697,0.48449,0.23925,1.16246,0.84003,0.00083665,0.00083665,0.00083665
 35,134.022,0.29217,0.30036,0.82236,0.74019,0.62576,0.5455,0.50497,0.23662,1.12651,0.8294,0.0008317,0.0008317,0.0008317
 36,137.857,0.30496,0.29963,0.82871,0.77746,0.61902,0.54974,0.50864,0.23983,1.13266,0.82649,0.00082675,0.00082675,0.00082675
 37,141.675,0.29652,0.28504,0.83423,0.76224,0.61966,0.54832,0.51254,0.23244,1.10152,0.82014,0.0008218,0.0008218,0.0008218
 38,145.544,0.29316,0.29352,0.82902,0.77191,0.59872,0.54528,0.48635,0.24638,1.15076,0.83673,0.00081685,0.00081685,0.00081685
 39,149.389,0.29163,0.29201,0.82565,0.77629,0.60478,0.53634,0.47925,0.2394,1.16222,0.83068,0.0008119,0.0008119,0.0008119
 40,153.23,0.28631,0.2953,0.82819,0.7648,0.59892,0.55013,0.50008,0.24854,1.16313,0.83741,0.00080695,0.00080695,0.00080695
 41,157.077,0.29419,0.29807,0.82685,0.75683,0.59863,0.54184,0.4959,0.2474,1.24443,0.83277,0.000802,0.000802,0.000802
 42,160.942,0.29756,0.28626,0.82306,0.75421,0.57682,0.54342,0.49727,0.2331,1.17565,0.82465,0.00079705,0.00079705,0.00079705
 43,164.81,0.2871,0.28713,0.81764,0.74553,0.59723,0.53672,0.50959,0.23344,1.1223,0.81151,0.0007921,0.0007921,0.0007921
 44,168.676,0.2771,0.29087,0.82059,0.76691,0.56335,0.54888,0.49446,0.24563,1.0691,0.81451,0.00078715,0.00078715,0.00078715
 45,172.514,0.29064,0.28331,0.82157,0.7249,0.59497,0.54521,0.50336,0.23342,1.0906,0.81361,0.0007822,0.0007822,0.0007822
 46,176.387,0.2885,0.28109,0.82382,0.72961,0.60313,0.54047,0.49478,0.25753,1.13492,0.82459,0.00077725,0.00077725,0.00077725
 47,180.253,0.29194,0.28328,0.82439,0.7383,0.58453,0.53567,0.49367,0.26183,1.13942,0.82808,0.0007723,0.0007723,0.0007723
 48,184.123,0.28052,0.27973,0.81483,0.74037,0.58509,0.53808,0.48299,0.26022,1.17242,0.82552,0.00076735,0.00076735,0.00076735
 49,187.962,0.2843,0.27279,0.82323,0.74705,0.5827,0.52754,0.45222,0.27769,1.08989,0.83224,0.0007624,0.0007624,0.0007624
 50,191.822,0.27583,0.27387,0.81463,0.77078,0.51457,0.52907,0.45994,0.26412,1.02878,0.83126,0.00075745,0.00075745,0.00075745
 51,195.695,0.28481,0.27865,0.8206,0.77605,0.52752,0.52099,0.45608,0.26288,1.09543,0.8257,0.0007525,0.0007525,0.0007525
--- a/fashion_yolo/finetuning_exp/finetuned_exp/yolov8m_finetuned_16class_cheated/args.yaml
+++ b/fashion_yolo/finetuning_exp/finetuned_exp/yolov8m_finetuned_16class_cheated/args.yaml
@ -0,0 +1,106 @@
 task: detect
 mode: train
 model: yolov8m.pt
 data: /home/cuuva/git/Detection_Experiment/fashion_yolo/finetuning_exp/finetuning_dataset.yaml
 epochs: 200
 time: null
 patience: 50
 batch: -1
 imgsz: 640
 save: true
 save_period: -1
 cache: false
 device: '0'
 workers: 8
 project: finetuned_exp
 name: yolov8m_finetuned_16class2
 exist_ok: false
 pretrained: true
 optimizer: AdamW
 verbose: true
 seed: 0
 deterministic: true
 single_cls: false
 rect: false
 cos_lr: false
 close_mosaic: 10
 resume: false
 amp: true
 fraction: 1.0
 profile: false
 freeze: null
 multi_scale: false
 compile: false
 overlap_mask: true
 mask_ratio: 4
 dropout: 0.0
 val: true
 split: val
 save_json: false
 conf: null
 iou: 0.7
 max_det: 300
 half: false
 dnn: false
 plots: true
 source: null
 vid_stride: 1
 stream_buffer: false
 visualize: false
 augment: false
 agnostic_nms: false
 classes: null
 retina_masks: false
 embed: null
 show: false
 save_frames: false
 save_txt: false
 save_conf: false
 save_crop: false
 show_labels: true
 show_conf: true
 show_boxes: true
 line_width: null
 format: torchscript
 keras: false
 optimize: false
 int8: false
 dynamic: false
 simplify: true
 opset: null
 workspace: null
 nms: false
 lr0: 0.001
 lrf: 0.01
 momentum: 0.937
 weight_decay: 0.0005
 warmup_epochs: 3.0
 warmup_momentum: 0.8
 warmup_bias_lr: 0.1
 box: 7.5
 cls: 0.5
 dfl: 1.5
 pose: 12.0
 kobj: 1.0
 nbs: 64
 hsv_h: 0.015
 hsv_s: 0.7
 hsv_v: 0.4
 degrees: 0.0
 translate: 0.1
 scale: 0.5
 shear: 0.0
 perspective: 0.0
 flipud: 0.0
 fliplr: 0.5
 bgr: 0.0
 mosaic: 1.0
 mixup: 0.0
 cutmix: 0.0
 copy_paste: 0.0
 copy_paste_mode: flip
 auto_augment: randaugment
 erasing: 0.4
 cfg: null
 tracker: botsort.yaml
 save_dir: /home/cuuva/git/Detection_Experiment/fashion_yolo/finetuning_exp/finetuned_exp/yolov8m_finetuned_16class2
--- a/fashion_yolo/finetuning_exp/finetuned_exp/yolov8m_finetuned_16class_cheated/results.csv
+++ b/fashion_yolo/finetuning_exp/finetuned_exp/yolov8m_finetuned_16class_cheated/results.csv
@ -0,0 +1,201 @@
 epoch,time,train/box_loss,train/cls_loss,train/dfl_loss,metrics/precision(B),metrics/recall(B),metrics/mAP50(B),metrics/mAP50-95(B),val/box_loss,val/cls_loss,val/dfl_loss,lr/pg0,lr/pg1,lr/pg2
 1,4.28543,1.65329,4.70774,1.93232,0.55379,0.62841,0.38518,0.26619,1.20824,2.31443,1.66836,0.09208,8e-05,8e-05
 2,8.54589,0.60909,1.19382,1.00586,0.56835,0.37769,0.317,0.23683,0.65806,2.90476,1.06487,0.0831692,0.000169158,0.000169158
 3,12.8001,0.51406,0.74646,0.91683,0.39877,0.45534,0.34991,0.27053,0.7152,1.47092,1.18814,0.0742574,0.000257426,0.000257426
 4,17.1085,0.50935,0.65683,0.90693,0.29185,0.68435,0.35795,0.28151,0.44693,2.14368,0.89714,0.0653448,0.000344802,0.000344802
 5,21.4268,0.49961,0.67146,0.90504,0.52561,0.53864,0.56972,0.45648,0.46344,1.63025,0.90896,0.0564313,0.000431288,0.000431288
 6,25.7708,0.47671,0.62084,0.89061,0.80651,0.56621,0.60681,0.55763,0.41707,1.67674,0.90167,0.0475169,0.000516883,0.000516883
 7,30.1356,0.4827,0.57954,0.89412,0.71877,0.6295,0.6163,0.50227,0.39606,2.18887,0.88696,0.0386016,0.000601586,0.000601586
 8,34.3853,0.50058,0.57879,0.90453,0.78319,0.64458,0.56321,0.52137,0.31404,1.18172,0.8507,0.0296854,0.000685398,0.000685398
 9,38.636,0.48863,0.58284,0.8889,0.54497,0.54017,0.38706,0.33093,0.5133,2.48353,0.92821,0.0207683,0.00076832,0.00076832
 10,42.885,0.51538,0.6252,0.91273,0.38568,0.61523,0.43533,0.36232,0.50516,1.49371,0.9343,0.0118504,0.00085035,0.00085035
 11,47.1388,0.51593,0.59906,0.90275,0.58358,0.46257,0.36506,0.31172,0.52138,1.42125,0.98901,0.00293149,0.00093149,0.00093149
 12,51.3864,0.53071,0.59235,0.90451,0.46905,0.45114,0.27714,0.22731,0.47681,1.51678,0.91254,0.00094555,0.00094555,0.00094555
 13,55.6696,0.51302,0.58747,0.90491,0.62926,0.39357,0.22728,0.20392,0.91744,6.46639,1.29086,0.0009406,0.0009406,0.0009406
 14,59.9581,0.49596,0.55809,0.89526,0.72468,0.25307,0.19564,0.16913,0.83195,4.17386,1.27834,0.00093565,0.00093565,0.00093565
 15,64.2461,0.47345,0.55318,0.89335,0.66879,0.40682,0.37083,0.27735,1.11479,2.11375,1.57138,0.0009307,0.0009307,0.0009307
 16,68.5313,0.47408,0.54535,0.88864,0.72928,0.27538,0.36783,0.27617,1.09593,2.60619,1.34331,0.00092575,0.00092575,0.00092575
 17,72.819,0.48671,0.55072,0.89224,0.74597,0.5832,0.58541,0.51619,0.47841,1.54962,0.92983,0.0009208,0.0009208,0.0009208
 18,77.0734,0.46808,0.53394,0.88852,0.64165,0.34351,0.3874,0.32198,0.73906,1.59411,1.28145,0.00091585,0.00091585,0.00091585
 19,81.3668,0.45283,0.53056,0.8899,0.67527,0.54147,0.50275,0.44812,0.427,0.99522,0.90487,0.0009109,0.0009109,0.0009109
 20,85.6547,0.46472,0.52678,0.89441,0.71214,0.67841,0.56492,0.52156,0.37245,0.75573,0.86661,0.00090595,0.00090595,0.00090595
 21,89.9468,0.46891,0.52417,0.88574,0.72736,0.66044,0.57811,0.48184,0.37352,0.78595,0.89237,0.000901,0.000901,0.000901
 22,94.2106,0.4465,0.49896,0.87858,0.74979,0.68561,0.61768,0.58342,0.34831,1.30751,0.86364,0.00089605,0.00089605,0.00089605
 23,98.5348,0.43933,0.49852,0.88394,0.85443,0.51932,0.61049,0.554,0.38463,1.20985,0.88358,0.0008911,0.0008911,0.0008911
 24,102.792,0.43604,0.50003,0.88906,0.86467,0.65568,0.68155,0.62936,0.35501,0.61947,0.85598,0.00088615,0.00088615,0.00088615
 25,107.124,0.44841,0.50766,0.88754,0.78848,0.59659,0.64362,0.59278,0.3624,0.89102,0.86272,0.0008812,0.0008812,0.0008812
 26,111.435,0.42763,0.49466,0.86851,0.72239,0.71822,0.60752,0.54966,0.39193,1.35293,0.89923,0.00087625,0.00087625,0.00087625
 27,115.736,0.43199,0.47031,0.87941,0.71628,0.66098,0.64998,0.59117,0.34432,1.0775,0.85677,0.0008713,0.0008713,0.0008713
 28,120.029,0.42158,0.47946,0.87222,0.82736,0.69015,0.63881,0.57956,0.3059,0.84559,0.83088,0.00086635,0.00086635,0.00086635
 29,124.301,0.42462,0.48392,0.86614,0.79084,0.68054,0.63945,0.60285,0.28541,0.74616,0.82473,0.0008614,0.0008614,0.0008614
 30,128.561,0.41487,0.46859,0.86652,0.8473,0.6185,0.6182,0.59573,0.2799,0.60409,0.82277,0.00085645,0.00085645,0.00085645
 31,132.837,0.42476,0.45112,0.87834,0.83242,0.66539,0.64273,0.57868,0.28277,0.63505,0.81473,0.0008515,0.0008515,0.0008515
 32,137.11,0.40256,0.42978,0.86946,0.86595,0.61132,0.65273,0.60448,0.29416,0.7586,0.81505,0.00084655,0.00084655,0.00084655
 33,141.421,0.42254,0.49085,0.88208,0.82789,0.63615,0.66188,0.61326,0.28153,0.62772,0.81144,0.0008416,0.0008416,0.0008416
 34,145.712,0.43286,0.48739,0.89036,0.83298,0.67386,0.66438,0.64412,0.27574,0.86448,0.80618,0.00083665,0.00083665,0.00083665
 35,150.078,0.40767,0.45055,0.865,0.87673,0.59848,0.68492,0.65016,0.28831,0.54603,0.82598,0.0008317,0.0008317,0.0008317
 36,154.433,0.40941,0.43927,0.87481,0.858,0.6278,0.68334,0.6453,0.26206,0.55396,0.80859,0.00082675,0.00082675,0.00082675
 37,158.716,0.39553,0.43671,0.86734,0.83746,0.68939,0.71626,0.66796,0.27281,0.61677,0.82392,0.0008218,0.0008218,0.0008218
 38,163.056,0.39026,0.4304,0.85655,0.84982,0.6324,0.69005,0.65319,0.28145,0.55189,0.83358,0.00081685,0.00081685,0.00081685
 39,167.328,0.39697,0.44514,0.86036,0.88586,0.68561,0.72929,0.66609,0.29897,0.54627,0.83789,0.0008119,0.0008119,0.0008119
 40,171.655,0.39548,0.4536,0.85825,0.56278,0.94735,0.81584,0.73417,0.26922,0.53472,0.82167,0.00080695,0.00080695,0.00080695
 41,176.019,0.40193,0.41767,0.86015,0.88176,0.6765,0.74373,0.69085,0.27001,0.57578,0.82596,0.000802,0.000802,0.000802
 42,180.299,0.37976,0.41144,0.8546,0.87817,0.65833,0.71454,0.67631,0.25682,0.47072,0.81671,0.00079705,0.00079705,0.00079705
 43,184.579,0.37324,0.41658,0.85218,0.89593,0.67386,0.71199,0.66064,0.28672,0.53076,0.82679,0.0007921,0.0007921,0.0007921
 44,188.872,0.3862,0.42587,0.85838,0.88063,0.67963,0.72944,0.69899,0.25952,0.4542,0.81391,0.00078715,0.00078715,0.00078715
 45,193.145,0.38377,0.41995,0.85917,0.92941,0.64112,0.72157,0.67033,0.27212,0.64078,0.82381,0.0007822,0.0007822,0.0007822
 46,197.452,0.37504,0.40806,0.85822,0.86094,0.70368,0.7171,0.67795,0.25614,0.74845,0.82338,0.00077725,0.00077725,0.00077725
 47,201.727,0.37447,0.41709,0.85656,0.84283,0.61864,0.71093,0.65883,0.2824,0.6009,0.82634,0.0007723,0.0007723,0.0007723
 48,206.039,0.3757,0.41362,0.85355,0.85214,0.67388,0.93103,0.81695,0.28495,0.49182,0.82113,0.00076735,0.00076735,0.00076735
 49,210.389,0.37336,0.42715,0.85708,0.55086,0.97008,0.94319,0.83678,0.28684,0.48025,0.82525,0.0007624,0.0007624,0.0007624
 50,214.759,0.36918,0.39176,0.8543,0.92588,0.66925,0.95187,0.86164,0.25244,0.44716,0.81859,0.00075745,0.00075745,0.00075745
 51,219.138,0.3539,0.39374,0.84924,0.9269,0.67267,0.82204,0.74372,0.28873,0.50146,0.82173,0.0007525,0.0007525,0.0007525
 52,223.437,0.37608,0.37839,0.8523,0.9374,0.66076,0.71568,0.65843,0.26263,0.42631,0.82319,0.00074755,0.00074755,0.00074755
 53,227.717,0.36571,0.39501,0.85776,0.92478,0.65303,0.72564,0.66866,0.26609,0.49754,0.82106,0.0007426,0.0007426,0.0007426
 54,231.999,0.368,0.38568,0.85594,0.85787,0.68877,0.69977,0.65519,0.23691,0.52451,0.81995,0.00073765,0.00073765,0.00073765
 55,236.317,0.37165,0.40803,0.85439,0.8035,0.67386,0.76173,0.70748,0.25569,0.44362,0.81317,0.0007327,0.0007327,0.0007327
 56,240.633,0.37852,0.40014,0.85273,0.90792,0.66397,0.76746,0.72622,0.2372,0.44036,0.81114,0.00072775,0.00072775,0.00072775
 57,244.946,0.35346,0.37178,0.84849,0.84354,0.66355,0.82738,0.77069,0.24262,0.49255,0.80392,0.0007228,0.0007228,0.0007228
 58,249.235,0.35021,0.36446,0.84335,0.86266,0.66098,0.74236,0.71792,0.23379,0.49133,0.80423,0.00071785,0.00071785,0.00071785
 59,253.51,0.35077,0.37176,0.85222,0.94838,0.66078,0.77509,0.71456,0.24736,0.37968,0.81331,0.0007129,0.0007129,0.0007129
 60,257.796,0.35622,0.3941,0.84949,0.9101,0.66228,0.75278,0.71534,0.24834,0.51509,0.79988,0.00070795,0.00070795,0.00070795
 61,262.114,0.34685,0.37241,0.84552,0.96372,0.64294,0.72076,0.68883,0.22526,0.38434,0.80634,0.000703,0.000703,0.000703
 62,266.44,0.34037,0.38093,0.84274,0.88242,0.652,0.70824,0.66667,0.26109,0.53399,0.83942,0.00069805,0.00069805,0.00069805
 63,270.75,0.33665,0.36925,0.8426,0.90719,0.68769,0.74511,0.70897,0.2406,0.45853,0.82805,0.0006931,0.0006931,0.0006931
 64,275.08,0.34441,0.38482,0.84215,0.94329,0.66758,0.821,0.7563,0.23285,0.449,0.81854,0.00068815,0.00068815,0.00068815
 65,279.367,0.34313,0.38034,0.84722,0.94161,0.68295,0.77084,0.73669,0.22787,0.40201,0.81282,0.0006832,0.0006832,0.0006832
 66,283.651,0.35207,0.36056,0.8525,0.87985,0.65064,0.68479,0.65578,0.2199,0.53037,0.81455,0.00067825,0.00067825,0.00067825
 67,287.954,0.33612,0.35198,0.83376,0.86645,0.67772,0.6995,0.68154,0.23403,0.47777,0.80969,0.0006733,0.0006733,0.0006733
 68,292.278,0.33159,0.36299,0.83649,0.90149,0.68364,0.74759,0.72412,0.22437,0.38223,0.81581,0.00066835,0.00066835,0.00066835
 69,296.604,0.34118,0.35332,0.84064,0.91715,0.68419,0.71379,0.68747,0.20496,0.43167,0.8038,0.0006634,0.0006634,0.0006634
 70,300.924,0.33177,0.34555,0.84348,0.96104,0.63415,0.70373,0.68437,0.21185,0.43304,0.80507,0.00065845,0.00065845,0.00065845
 71,305.245,0.32953,0.35183,0.8363,0.93873,0.65744,0.706,0.68622,0.23041,0.47325,0.80616,0.0006535,0.0006535,0.0006535
 72,309.533,0.32,0.33748,0.83928,0.89685,0.70764,0.71115,0.67818,0.25043,0.44521,0.82634,0.00064855,0.00064855,0.00064855
 73,313.831,0.34052,0.34484,0.85154,0.96019,0.64363,0.73122,0.70664,0.228,0.41417,0.82115,0.0006436,0.0006436,0.0006436
 74,318.107,0.32634,0.35581,0.83818,0.92189,0.67197,0.74703,0.71628,0.22408,0.47403,0.8108,0.00063865,0.00063865,0.00063865
 75,322.441,0.32855,0.36429,0.85078,0.93739,0.70582,0.74014,0.71116,0.21411,0.38648,0.81334,0.0006337,0.0006337,0.0006337
 76,326.775,0.32295,0.34231,0.83829,0.95709,0.67799,0.95758,0.90544,0.21061,0.34109,0.81508,0.00062875,0.00062875,0.00062875
 77,331.157,0.32076,0.33621,0.83241,0.74961,0.95644,0.96407,0.89205,0.23334,0.39555,0.81005,0.0006238,0.0006238,0.0006238
 78,335.476,0.32787,0.33604,0.83646,0.77449,0.96969,0.96826,0.89719,0.20603,0.3486,0.8054,0.00061885,0.00061885,0.00061885
 79,339.784,0.33131,0.34025,0.83239,0.93983,0.84039,0.96767,0.90409,0.21413,0.39012,0.81562,0.0006139,0.0006139,0.0006139
 80,344.066,0.32582,0.33691,0.84696,0.82592,0.95644,0.96471,0.9178,0.21615,0.3521,0.81822,0.00060895,0.00060895,0.00060895
 81,348.411,0.32543,0.32518,0.83818,0.8178,0.95863,0.96809,0.87722,0.23237,0.35475,0.82237,0.000604,0.000604,0.000604
 82,352.729,0.31963,0.31646,0.83792,0.70791,0.96045,0.95647,0.85567,0.23326,0.41919,0.81455,0.00059905,0.00059905,0.00059905
 83,357.048,0.32231,0.32536,0.83148,0.70971,0.97008,0.96151,0.91001,0.21876,0.39028,0.81159,0.0005941,0.0005941,0.0005941
 84,361.391,0.32808,0.33879,0.84383,0.65238,0.97008,0.96419,0.93221,0.21162,0.3588,0.81303,0.00058915,0.00058915,0.00058915
 85,365.746,0.31537,0.32334,0.83886,0.79425,0.963,0.96644,0.89883,0.22393,0.34411,0.81364,0.0005842,0.0005842,0.0005842
 86,370.03,0.32068,0.31532,0.83454,0.56245,0.95379,0.96666,0.88574,0.22058,0.38905,0.81247,0.00057925,0.00057925,0.00057925
 87,374.311,0.32101,0.31908,0.83926,0.76162,0.72462,0.95039,0.88236,0.23339,0.36622,0.81958,0.0005743,0.0005743,0.0005743
 88,378.605,0.30519,0.30833,0.83157,0.92783,0.65799,0.96228,0.89174,0.20367,0.3928,0.80613,0.00056935,0.00056935,0.00056935
 89,382.927,0.30314,0.30569,0.82607,0.95232,0.68992,0.95813,0.90097,0.208,0.31056,0.80571,0.0005644,0.0005644,0.0005644
 90,387.242,0.32365,0.32198,0.85116,0.65009,0.97008,0.96249,0.90673,0.20781,0.32799,0.80194,0.00055945,0.00055945,0.00055945
 91,391.542,0.323,0.34086,0.84289,0.9539,0.71553,0.96769,0.90124,0.2139,0.32589,0.80636,0.0005545,0.0005545,0.0005545
 92,395.837,0.30533,0.30191,0.83812,0.96256,0.69015,0.84339,0.78966,0.20946,0.36547,0.8107,0.00054955,0.00054955,0.00054955
 93,400.127,0.30469,0.29959,0.82611,0.95384,0.70847,0.96747,0.89782,0.19865,0.27637,0.79993,0.0005446,0.0005446,0.0005446
 94,404.44,0.31573,0.31115,0.83594,0.75591,0.97462,0.9579,0.88906,0.20908,0.31405,0.8044,0.00053965,0.00053965,0.00053965
 95,408.732,0.3166,0.32196,0.83967,0.97751,0.67305,0.95825,0.88757,0.19285,0.3869,0.80232,0.0005347,0.0005347,0.0005347
 96,413.065,0.31002,0.31908,0.83736,0.93007,0.72008,0.96417,0.90358,0.19903,0.3106,0.80161,0.00052975,0.00052975,0.00052975
 97,417.353,0.30327,0.30042,0.83168,0.92678,0.66543,0.83016,0.80649,0.18801,0.38408,0.79687,0.0005248,0.0005248,0.0005248
 98,421.689,0.3016,0.29055,0.83714,0.95264,0.64414,0.83419,0.78518,0.18983,0.32975,0.79538,0.00051985,0.00051985,0.00051985
 99,425.977,0.30319,0.30273,0.83317,0.82212,0.95379,0.96095,0.89586,0.20004,0.3866,0.8043,0.0005149,0.0005149,0.0005149
 100,430.266,0.30388,0.29825,0.83257,0.90006,0.68646,0.95453,0.86999,0.2074,0.41483,0.81063,0.00050995,0.00050995,0.00050995
 101,434.586,0.29649,0.29365,0.82556,0.84245,0.72431,0.96205,0.91564,0.18243,0.33307,0.79535,0.000505,0.000505,0.000505
 102,438.876,0.29276,0.29682,0.82893,0.83612,0.97462,0.97218,0.91231,0.18732,0.28054,0.79735,0.00050005,0.00050005,0.00050005
 103,443.192,0.30473,0.29691,0.84479,0.84583,0.97008,0.97264,0.93131,0.2076,0.28394,0.7997,0.0004951,0.0004951,0.0004951
 104,447.524,0.29763,0.29922,0.82446,0.90012,0.95876,0.97376,0.93627,0.18205,0.28868,0.79839,0.00049015,0.00049015,0.00049015
 105,451.906,0.29631,0.29083,0.83144,0.84888,0.96553,0.97155,0.93421,0.18822,0.2643,0.79615,0.0004852,0.0004852,0.0004852
 106,456.237,0.29928,0.30125,0.8366,0.67106,0.97462,0.97338,0.90661,0.18997,0.24255,0.79768,0.00048025,0.00048025,0.00048025
 107,460.522,0.29511,0.29164,0.83173,0.65779,0.97462,0.97187,0.90609,0.18033,0.29786,0.79276,0.0004753,0.0004753,0.0004753
 108,464.818,0.2789,0.29004,0.82999,0.97567,0.71626,0.97433,0.90612,0.18339,0.25326,0.79549,0.00047035,0.00047035,0.00047035
 109,469.11,0.28188,0.28147,0.82904,0.70615,0.97462,0.97366,0.94265,0.18238,0.25242,0.79368,0.0004654,0.0004654,0.0004654
 110,473.483,0.29345,0.28236,0.826,0.8821,0.96553,0.97163,0.95003,0.19111,0.33093,0.79519,0.00046045,0.00046045,0.00046045
 111,477.837,0.28831,0.27292,0.82335,0.81456,0.96553,0.96544,0.95911,0.17687,0.28484,0.7967,0.0004555,0.0004555,0.0004555
 112,482.191,0.29585,0.29073,0.83274,0.84611,0.93628,0.95716,0.92482,0.18754,0.29666,0.80057,0.00045055,0.00045055,0.00045055
 113,486.502,0.2982,0.29575,0.8267,0.85615,0.95553,0.95728,0.9483,0.19339,0.39318,0.80616,0.0004456,0.0004456,0.0004456
 114,490.79,0.28435,0.28632,0.81795,0.93313,0.95189,0.97191,0.96235,0.17436,0.29192,0.79601,0.00044065,0.00044065,0.00044065
 115,495.182,0.29767,0.27911,0.83392,0.8966,0.95352,0.96931,0.96098,0.18414,0.29926,0.79933,0.0004357,0.0004357,0.0004357
 116,499.51,0.27586,0.27774,0.82536,0.89243,0.96466,0.96956,0.96463,0.18766,0.3335,0.79988,0.00043075,0.00043075,0.00043075
 117,503.871,0.28466,0.26873,0.83141,0.77551,0.97008,0.96877,0.93379,0.18778,0.31004,0.79862,0.0004258,0.0004258,0.0004258
 118,508.204,0.27684,0.27792,0.82939,0.87693,0.97008,0.96288,0.90259,0.1863,0.45805,0.79516,0.00042085,0.00042085,0.00042085
 119,512.503,0.27553,0.27752,0.82132,0.92584,0.96098,0.97189,0.9063,0.18499,0.25587,0.79907,0.0004159,0.0004159,0.0004159
 120,516.785,0.27542,0.26814,0.83313,0.87093,0.97008,0.97053,0.93002,0.1754,0.29144,0.7979,0.00041095,0.00041095,0.00041095
 121,521.101,0.27132,0.2561,0.81791,0.92567,0.96098,0.97137,0.91018,0.19108,0.29594,0.79978,0.000406,0.000406,0.000406
 122,525.422,0.26509,0.25866,0.81853,0.90199,0.96098,0.97352,0.90916,0.18136,0.24954,0.79734,0.00040105,0.00040105,0.00040105
 123,529.718,0.28063,0.26065,0.81698,0.81092,0.97462,0.9727,0.91217,0.17841,0.24652,0.79399,0.0003961,0.0003961,0.0003961
 124,534.013,0.26959,0.26148,0.83041,0.84504,0.97462,0.96756,0.90897,0.17546,0.23145,0.79512,0.00039115,0.00039115,0.00039115
 125,538.299,0.2785,0.26648,0.82962,0.86591,0.97462,0.97436,0.9154,0.16593,0.19827,0.79506,0.0003862,0.0003862,0.0003862
 126,542.626,0.27278,0.25455,0.82567,0.8759,0.97462,0.97258,0.90767,0.16722,0.21979,0.79325,0.00038125,0.00038125,0.00038125
 127,546.945,0.26655,0.24821,0.81653,0.8409,0.97462,0.97292,0.93896,0.16524,0.21884,0.79287,0.0003763,0.0003763,0.0003763
 128,551.242,0.26137,0.24664,0.82216,0.90076,0.96868,0.9695,0.92962,0.18153,0.20633,0.79503,0.00037135,0.00037135,0.00037135
 129,555.54,0.26889,0.25579,0.8275,0.89604,0.97008,0.97382,0.94517,0.16841,0.20242,0.7942,0.0003664,0.0003664,0.0003664
 130,559.823,0.27153,0.25214,0.82088,0.95901,0.96073,0.97636,0.96652,0.16974,0.19452,0.79431,0.00036145,0.00036145,0.00036145
 131,564.188,0.27584,0.26728,0.82943,0.9668,0.95836,0.97728,0.94103,0.17023,0.19748,0.79211,0.0003565,0.0003565,0.0003565
 132,568.512,0.26903,0.25414,0.83224,0.90101,0.97008,0.97383,0.9695,0.16216,0.22065,0.79097,0.00035155,0.00035155,0.00035155
 133,572.901,0.25954,0.25001,0.81623,0.92189,0.95644,0.96823,0.95222,0.16911,0.26631,0.79164,0.0003466,0.0003466,0.0003466
 134,577.225,0.25502,0.25049,0.82106,0.94679,0.95644,0.97495,0.93015,0.1677,0.2006,0.79421,0.00034165,0.00034165,0.00034165
 135,581.51,0.26729,0.25724,0.82081,0.95692,0.95613,0.97552,0.96289,0.17356,0.1896,0.79329,0.0003367,0.0003367,0.0003367
 136,585.795,0.26559,0.25837,0.82502,0.93999,0.96553,0.97528,0.93348,0.16785,0.17389,0.79263,0.00033175,0.00033175,0.00033175
 137,590.084,0.26478,0.25906,0.8187,0.96728,0.95443,0.9742,0.93189,0.15991,0.3405,0.79021,0.0003268,0.0003268,0.0003268
 138,594.425,0.26044,0.23131,0.81842,0.97404,0.93774,0.97656,0.93271,0.16447,0.21099,0.79091,0.00032185,0.00032185,0.00032185
 139,598.754,0.25589,0.23997,0.8166,0.96069,0.95973,0.97645,0.94019,0.16396,0.16737,0.78957,0.0003169,0.0003169,0.0003169
 140,603.052,0.26577,0.23646,0.82005,0.97855,0.96072,0.9767,0.93968,0.17007,0.16758,0.78916,0.00031195,0.00031195,0.00031195
 141,607.374,0.26244,0.23761,0.8199,0.94171,0.97008,0.97518,0.93338,0.16998,0.15943,0.78875,0.000307,0.000307,0.000307
 142,611.683,0.25333,0.22862,0.81463,0.93562,0.97008,0.97451,0.96451,0.16112,0.16873,0.78851,0.00030205,0.00030205,0.00030205
 143,615.969,0.24801,0.24092,0.81858,0.97665,0.95973,0.97608,0.94219,0.16409,0.15063,0.7881,0.0002971,0.0002971,0.0002971
 144,620.262,0.26638,0.24675,0.82719,0.9337,0.97008,0.97679,0.94269,0.16296,0.18705,0.79056,0.00029215,0.00029215,0.00029215
 145,624.585,0.25794,0.24589,0.81336,0.90296,0.97008,0.97718,0.94789,0.15475,0.16331,0.78891,0.0002872,0.0002872,0.0002872
 146,628.92,0.25254,0.24709,0.82283,0.96564,0.96098,0.97536,0.96445,0.15514,0.16403,0.78554,0.00028225,0.00028225,0.00028225
 147,633.212,0.25929,0.23812,0.82427,0.96594,0.96416,0.97597,0.96736,0.1603,0.15395,0.78752,0.0002773,0.0002773,0.0002773
 148,637.509,0.2536,0.23606,0.81486,0.97062,0.96403,0.97667,0.94672,0.15692,0.16712,0.78897,0.00027235,0.00027235,0.00027235
 149,641.804,0.24436,0.2252,0.81162,0.97313,0.96693,0.97619,0.94203,0.15185,0.17111,0.78554,0.0002674,0.0002674,0.0002674
 150,646.093,0.25437,0.23935,0.82815,0.97207,0.96805,0.97809,0.94901,0.14466,0.1587,0.77975,0.00026245,0.00026245,0.00026245
 151,650.341,0.25438,0.24148,0.82491,0.95852,0.9656,0.9782,0.94813,0.14126,0.15164,0.78076,0.0002575,0.0002575,0.0002575
 152,654.629,0.25356,0.22954,0.82656,0.95484,0.96553,0.97668,0.9717,0.14693,0.16615,0.7831,0.00025255,0.00025255,0.00025255
 153,659.013,0.24174,0.21109,0.82237,0.98303,0.95495,0.97584,0.93056,0.15264,0.16628,0.78287,0.0002476,0.0002476,0.0002476
 154,663.339,0.24187,0.22267,0.8136,0.97049,0.96052,0.973,0.93441,0.15029,0.18946,0.78053,0.00024265,0.00024265,0.00024265
 155,667.645,0.24514,0.21244,0.81664,0.96083,0.96175,0.97359,0.93551,0.1552,0.20072,0.78392,0.0002377,0.0002377,0.0002377
 156,671.947,0.24136,0.22267,0.81732,0.97031,0.95858,0.97369,0.9302,0.14873,0.16372,0.78197,0.00023275,0.00023275,0.00023275
 157,676.234,0.24092,0.21925,0.8157,0.97283,0.96872,0.97365,0.93757,0.14744,0.14907,0.78168,0.0002278,0.0002278,0.0002278
 158,680.566,0.23119,0.21237,0.81421,0.95689,0.97008,0.97325,0.96185,0.15194,0.15399,0.78306,0.00022285,0.00022285,0.00022285
 159,684.863,0.23864,0.21986,0.81152,0.94659,0.97008,0.97796,0.96675,0.14937,0.15875,0.78363,0.0002179,0.0002179,0.0002179
 160,689.196,0.23941,0.21688,0.81236,0.97344,0.96098,0.97858,0.96591,0.14238,0.15135,0.78583,0.00021295,0.00021295,0.00021295
 161,693.482,0.23887,0.21348,0.8104,0.98114,0.96098,0.9792,0.96705,0.14299,0.14865,0.78734,0.000208,0.000208,0.000208
 162,697.773,0.24129,0.22085,0.81968,0.97226,0.95379,0.97929,0.96816,0.14543,0.1689,0.78859,0.00020305,0.00020305,0.00020305
 163,702.094,0.23091,0.19975,0.80896,0.98381,0.95692,0.97922,0.96797,0.1414,0.15383,0.78656,0.0001981,0.0001981,0.0001981
 164,706.41,0.23283,0.20863,0.8117,0.98479,0.95609,0.97984,0.96134,0.14242,0.15036,0.78555,0.00019315,0.00019315,0.00019315
 165,710.73,0.22825,0.209,0.80686,0.98698,0.95644,0.97951,0.96741,0.13915,0.1512,0.78437,0.0001882,0.0001882,0.0001882
 166,715.059,0.23329,0.21597,0.80918,0.98638,0.95758,0.97757,0.94845,0.14553,0.15108,0.78464,0.00018325,0.00018325,0.00018325
 167,719.382,0.24392,0.20815,0.81207,0.98704,0.96563,0.97833,0.94933,0.14141,0.13066,0.78342,0.0001783,0.0001783,0.0001783
 168,723.697,0.23426,0.20962,0.81154,0.98724,0.96803,0.97924,0.9752,0.14035,0.11908,0.78182,0.00017335,0.00017335,0.00017335
 169,728.067,0.22972,0.20038,0.81226,0.98839,0.96443,0.98075,0.96908,0.13934,0.13037,0.78139,0.0001684,0.0001684,0.0001684
 170,732.366,0.22821,0.20397,0.81283,0.98645,0.96163,0.98178,0.96961,0.13915,0.12869,0.78066,0.00016345,0.00016345,0.00016345
 171,736.651,0.2318,0.20327,0.81026,0.98765,0.96701,0.98181,0.97026,0.14202,0.1267,0.78175,0.0001585,0.0001585,0.0001585
 172,740.954,0.22318,0.19867,0.80688,0.98381,0.96553,0.98021,0.97133,0.13056,0.12134,0.77829,0.00015355,0.00015355,0.00015355
 173,745.296,0.23076,0.19951,0.8109,0.98667,0.96869,0.97958,0.97501,0.12693,0.10893,0.77733,0.0001486,0.0001486,0.0001486
 174,749.622,0.23074,0.20295,0.81441,0.98812,0.96462,0.97972,0.97518,0.12908,0.10939,0.77696,0.00014365,0.00014365,0.00014365
 175,753.956,0.22397,0.19379,0.81298,0.98634,0.96492,0.98004,0.97562,0.13043,0.11054,0.77545,0.0001387,0.0001387,0.0001387
 176,758.354,0.22645,0.20291,0.81623,0.98819,0.96635,0.98259,0.97828,0.1277,0.11343,0.77592,0.00013375,0.00013375,0.00013375
 177,762.712,0.22742,0.19383,0.81473,0.98874,0.96452,0.98188,0.97757,0.13073,0.11041,0.77685,0.0001288,0.0001288,0.0001288
 178,767.024,0.21928,0.19457,0.80982,0.98891,0.96104,0.98209,0.9778,0.12701,0.11237,0.77573,0.00012385,0.00012385,0.00012385
 179,771.303,0.22433,0.19704,0.80745,0.98842,0.95599,0.98282,0.97858,0.12645,0.11494,0.7749,0.0001189,0.0001189,0.0001189
 180,775.687,0.20776,0.18235,0.80641,0.98888,0.96556,0.98206,0.97763,0.13294,0.1055,0.77661,0.00011395,0.00011395,0.00011395
 181,779.981,0.2189,0.19182,0.81262,0.98881,0.96707,0.98142,0.97737,0.12598,0.09987,0.77617,0.000109,0.000109,0.000109
 182,784.315,0.22136,0.19491,0.81348,0.98736,0.96564,0.9809,0.97701,0.12663,0.10497,0.77466,0.00010405,0.00010405,0.00010405
 183,788.612,0.21699,0.18694,0.80382,0.98633,0.9668,0.98049,0.97718,0.12245,0.11063,0.7726,9.91e-05,9.91e-05,9.91e-05
 184,792.903,0.21566,0.18789,0.81539,0.98744,0.96622,0.98203,0.97888,0.12273,0.10585,0.77174,9.415e-05,9.415e-05,9.415e-05
 185,797.252,0.2225,0.20242,0.81146,0.98731,0.96639,0.98203,0.95382,0.12596,0.10258,0.77216,8.92e-05,8.92e-05,8.92e-05
 186,801.573,0.21533,0.18832,0.80837,0.98818,0.96335,0.982,0.95612,0.1204,0.10072,0.77252,8.425e-05,8.425e-05,8.425e-05
 187,805.922,0.21941,0.19119,0.81246,0.98891,0.9638,0.98199,0.98098,0.11741,0.10284,0.77271,7.93e-05,7.93e-05,7.93e-05
 188,810.32,0.20607,0.1758,0.80048,0.98916,0.96044,0.9834,0.98239,0.11773,0.10092,0.77243,7.435e-05,7.435e-05,7.435e-05
 189,814.679,0.21303,0.18249,0.81172,0.98792,0.96252,0.9809,0.97987,0.12036,0.10688,0.77235,6.94e-05,6.94e-05,6.94e-05
 190,819.005,0.21523,0.19032,0.80994,0.98535,0.96667,0.98006,0.97903,0.12079,0.11722,0.77234,6.445e-05,6.445e-05,6.445e-05
 191,824.131,0.17859,0.15079,0.77278,0.98483,0.96815,0.98004,0.97557,0.12125,0.11329,0.7713,5.95e-05,5.95e-05,5.95e-05
 192,828.427,0.16988,0.15478,0.77117,0.98842,0.96693,0.9808,0.97639,0.12617,0.12272,0.77194,5.455e-05,5.455e-05,5.455e-05
 193,832.715,0.16965,0.13842,0.77254,0.98698,0.96907,0.97997,0.9755,0.1252,0.12012,0.77159,4.96e-05,4.96e-05,4.96e-05
 194,837.009,0.16504,0.13941,0.76016,0.98632,0.97008,0.97906,0.97474,0.12319,0.11815,0.76822,4.465e-05,4.465e-05,4.465e-05
 195,841.306,0.17513,0.14421,0.76998,0.98673,0.97462,0.97881,0.97443,0.12223,0.11097,0.76854,3.97e-05,3.97e-05,3.97e-05
 196,845.603,0.16749,0.13142,0.76763,0.98737,0.97375,0.9791,0.96817,0.12256,0.10862,0.76863,3.475e-05,3.475e-05,3.475e-05
 197,849.94,0.16698,0.13151,0.76674,0.98714,0.97462,0.97936,0.96888,0.12197,0.11223,0.76827,2.98e-05,2.98e-05,2.98e-05
 198,854.255,0.16759,0.13356,0.75944,0.98635,0.97462,0.97938,0.96875,0.12325,0.11181,0.76908,2.485e-05,2.485e-05,2.485e-05
 199,858.569,0.16581,0.13187,0.7688,0.98672,0.97462,0.97935,0.97535,0.12359,0.11041,0.76875,1.99e-05,1.99e-05,1.99e-05
 200,862.844,0.16452,0.13386,0.77479,0.98704,0.97462,0.97935,0.97535,0.12094,0.11219,0.76702,1.495e-05,1.495e-05,1.495e-05
--- a/fashion_yolo/finetuning_exp/finetuned_exp/yolov8m_finetuned_4class/args.yaml
+++ b/fashion_yolo/finetuning_exp/finetuned_exp/yolov8m_finetuned_4class/args.yaml
@ -0,0 +1,106 @@
 task: detect
 mode: train
 model: yolov8m.pt
 data: /home/cuuva/git/Detection_Experiment/fashion_yolo/finetuning_exp/finetuning_dataset.yaml
 epochs: 200
 time: null
 patience: 50
 batch: -1
 imgsz: 640
 save: true
 save_period: -1
 cache: false
 device: '0'
 workers: 8
 project: finetuned_exp
 name: yolov8m_finetuned_4class
 exist_ok: false
 pretrained: true
 optimizer: AdamW
 verbose: true
 seed: 0
 deterministic: true
 single_cls: false
 rect: false
 cos_lr: false
 close_mosaic: 10
 resume: false
 amp: true
 fraction: 1.0
 profile: false
 freeze: null
 multi_scale: false
 compile: false
 overlap_mask: true
 mask_ratio: 4
 dropout: 0.0
 val: true
 split: val
 save_json: false
 conf: null
 iou: 0.7
 max_det: 300
 half: false
 dnn: false
 plots: true
 source: null
 vid_stride: 1
 stream_buffer: false
 visualize: false
 augment: false
 agnostic_nms: false
 classes: null
 retina_masks: false
 embed: null
 show: false
 save_frames: false
 save_txt: false
 save_conf: false
 save_crop: false
 show_labels: true
 show_conf: true
 show_boxes: true
 line_width: null
 format: torchscript
 keras: false
 optimize: false
 int8: false
 dynamic: false
 simplify: true
 opset: null
 workspace: null
 nms: false
 lr0: 0.001
 lrf: 0.01
 momentum: 0.937
 weight_decay: 0.0005
 warmup_epochs: 3.0
 warmup_momentum: 0.8
 warmup_bias_lr: 0.1
 box: 7.5
 cls: 0.5
 dfl: 1.5
 pose: 12.0
 kobj: 1.0
 nbs: 64
 hsv_h: 0.015
 hsv_s: 0.7
 hsv_v: 0.4
 degrees: 0.0
 translate: 0.1
 scale: 0.5
 shear: 0.0
 perspective: 0.0
 flipud: 0.0
 fliplr: 0.5
 bgr: 0.0
 mosaic: 1.0
 mixup: 0.0
 cutmix: 0.0
 copy_paste: 0.0
 copy_paste_mode: flip
 auto_augment: randaugment
 erasing: 0.4
 cfg: null
 tracker: botsort.yaml
 save_dir: /home/cuuva/git/Detection_Experiment/fashion_yolo/finetuning_exp/finetuned_exp/yolov8m_finetuned_4class
--- a/fashion_yolo/finetuning_exp/finetuned_exp/yolov8m_finetuned_4class/results.csv
+++ b/fashion_yolo/finetuning_exp/finetuned_exp/yolov8m_finetuned_4class/results.csv
@ -0,0 +1,96 @@
 epoch,time,train/box_loss,train/cls_loss,train/dfl_loss,metrics/precision(B),metrics/recall(B),metrics/mAP50(B),metrics/mAP50-95(B),val/box_loss,val/cls_loss,val/dfl_loss,lr/pg0,lr/pg1,lr/pg2
 1,3.80278,1.48177,3.17142,1.82928,0.25935,0.63576,0.30094,0.25181,0.62926,9.19986,1.11815,0.08911,0.00011,0.00011
 2,7.56982,0.55462,0.81473,0.96504,0.51633,0.4058,0.18231,0.13309,0.98447,15.2608,1.39894,0.0772289,0.000228861,0.000228861
 3,11.35,0.49695,0.58671,0.91498,0.70773,0.63669,0.45707,0.39279,0.54234,1.05237,1.10085,0.0653465,0.000346535,0.000346535
 4,15.1889,0.49015,0.52583,0.91273,0.66188,0.63285,0.62662,0.56882,0.50988,1.39298,1.07138,0.053463,0.00046302,0.00046302
 5,19.0226,0.48089,0.51416,0.89657,0.69854,0.43478,0.48628,0.45263,0.41288,0.92681,1.0087,0.0415783,0.000578318,0.000578318
 6,22.9005,0.46976,0.54136,0.8959,0.36166,0.657,0.53062,0.48675,0.4253,0.96457,0.98522,0.0296924,0.000692427,0.000692427
 7,26.6362,0.43486,0.47627,0.88111,0.35734,0.907,0.42712,0.36855,0.48076,0.82531,1.02288,0.0178053,0.000805349,0.000805349
 8,30.3682,0.44514,0.46848,0.8846,0.37053,0.907,0.49135,0.41429,0.46108,0.79011,1.05954,0.00591708,0.000917082,0.000917082
 9,34.108,0.46561,0.49901,0.89686,0.69392,0.43478,0.42254,0.38327,0.44252,1.19051,1.0678,0.0009604,0.0009604,0.0009604
 10,37.8413,0.45597,0.47897,0.90645,0.42684,0.82367,0.56096,0.42621,0.44741,0.82823,1.08209,0.00095545,0.00095545,0.00095545
 11,41.6185,0.43596,0.46752,0.8799,0.77798,0.54589,0.60871,0.52497,0.43424,1.10056,1.03882,0.0009505,0.0009505,0.0009505
 12,45.3863,0.44031,0.46944,0.89503,0.76054,0.54589,0.53911,0.4219,0.47708,1.23622,1.08546,0.00094555,0.00094555,0.00094555
 13,49.2739,0.41508,0.4346,0.87447,0.41092,0.53682,0.51649,0.37632,0.45856,1.02469,1.06513,0.0009406,0.0009406,0.0009406
 14,53.0458,0.43565,0.44496,0.88068,0.36195,0.907,0.44093,0.39129,0.42948,0.81733,1.01262,0.00093565,0.00093565,0.00093565
 15,56.7821,0.42579,0.43487,0.88003,0.69472,0.657,0.43711,0.38686,0.49384,0.75617,1.0197,0.0009307,0.0009307,0.0009307
 16,60.5548,0.41614,0.45328,0.88708,0.73104,0.54589,0.51555,0.47644,0.49186,0.84718,0.9681,0.00092575,0.00092575,0.00092575
 17,64.3305,0.42171,0.42406,0.87461,0.7152,0.657,0.51313,0.46076,0.41855,0.715,0.97914,0.0009208,0.0009208,0.0009208
 18,68.1071,0.39919,0.40856,0.87174,0.69537,0.657,0.4992,0.44523,0.41394,0.79109,1.00867,0.00091585,0.00091585,0.00091585
 19,71.8535,0.39633,0.41433,0.86936,0.7401,0.63507,0.52952,0.47033,0.44341,0.65096,1.05446,0.0009109,0.0009109,0.0009109
 20,75.7038,0.40955,0.41822,0.88382,0.3939,0.85811,0.55513,0.47512,0.47293,0.81105,1.01416,0.00090595,0.00090595,0.00090595
 21,79.4863,0.39765,0.4193,0.86768,0.85051,0.42512,0.51614,0.44714,0.43829,0.9297,1.02766,0.000901,0.000901,0.000901
 22,83.2649,0.41389,0.42716,0.87677,0.69026,0.657,0.38712,0.35471,0.46728,0.77916,1.00062,0.00089605,0.00089605,0.00089605
 23,87.0201,0.40067,0.41106,0.87116,0.2922,0.86353,0.33118,0.29899,0.44699,1.27393,0.99676,0.0008911,0.0008911,0.0008911
 24,90.7882,0.41464,0.40279,0.8931,0.4414,0.51812,0.51282,0.43199,0.44852,0.88834,0.99389,0.00088615,0.00088615,0.00088615
 25,94.5736,0.3961,0.40144,0.87261,0.60054,0.82949,0.59507,0.4645,0.54252,0.94309,1.02617,0.0008812,0.0008812,0.0008812
 26,98.3608,0.38697,0.40143,0.86506,0.89368,0.54589,0.60658,0.54728,0.39207,0.59533,1.00538,0.00087625,0.00087625,0.00087625
 27,102.125,0.38916,0.40208,0.86755,0.3594,0.79589,0.55598,0.4948,0.38876,0.70555,0.9951,0.0008713,0.0008713,0.0008713
 28,105.987,0.37283,0.38684,0.86505,0.50015,0.85807,0.63872,0.53171,0.40371,0.64451,0.95713,0.00086635,0.00086635,0.00086635
 29,109.772,0.37602,0.38522,0.86444,0.49503,0.76349,0.64418,0.52735,0.37297,0.62525,0.9399,0.0008614,0.0008614,0.0008614
 30,113.524,0.37941,0.39047,0.8629,0.58795,0.7795,0.72367,0.63006,0.40849,0.57486,0.93323,0.00085645,0.00085645,0.00085645
 31,117.348,0.36408,0.37301,0.85983,0.41137,0.74034,0.58321,0.53515,0.3874,0.55179,0.93679,0.0008515,0.0008515,0.0008515
 32,121.126,0.37281,0.39046,0.85515,0.46201,0.82367,0.76309,0.63746,0.40881,0.51949,0.93422,0.00084655,0.00084655,0.00084655
 33,124.994,0.37661,0.37432,0.86613,0.53385,0.71256,0.71863,0.61991,0.41477,0.56341,0.94054,0.0008416,0.0008416,0.0008416
 34,128.849,0.36381,0.37272,0.85359,0.5407,0.76812,0.59987,0.51754,0.41479,0.76058,0.96833,0.00083665,0.00083665,0.00083665
 35,132.605,0.35963,0.38002,0.84906,0.54031,0.71256,0.5602,0.48774,0.41972,0.64676,1.02186,0.0008317,0.0008317,0.0008317
 36,136.361,0.36186,0.36353,0.85972,0.46755,0.68462,0.52039,0.43733,0.38765,0.68662,0.97133,0.00082675,0.00082675,0.00082675
 37,140.149,0.36793,0.37264,0.85513,0.63276,0.51812,0.56649,0.48905,0.38006,0.58723,0.94587,0.0008218,0.0008218,0.0008218
 38,143.936,0.36812,0.36672,0.86243,0.68112,0.70373,0.69821,0.6413,0.38211,0.62817,0.93824,0.00081685,0.00081685,0.00081685
 39,147.762,0.37858,0.37703,0.86357,0.57862,0.82367,0.71612,0.64133,0.40309,0.56662,1.00399,0.0008119,0.0008119,0.0008119
 40,151.584,0.3701,0.36609,0.85594,0.66083,0.77307,0.58969,0.52414,0.4089,0.63781,1.03948,0.00080695,0.00080695,0.00080695
 41,155.498,0.34569,0.35724,0.84426,0.67508,0.79442,0.5855,0.51378,0.41634,0.65617,1.0468,0.000802,0.000802,0.000802
 42,159.281,0.34801,0.35684,0.85775,0.63549,0.86167,0.63313,0.54981,0.43033,0.64608,1.03947,0.00079705,0.00079705,0.00079705
 43,163.069,0.35937,0.35179,0.86168,0.50612,0.84047,0.60053,0.54722,0.43482,0.54382,1.02906,0.0007921,0.0007921,0.0007921
 44,166.822,0.34928,0.34233,0.84939,0.55504,0.97767,0.75411,0.63998,0.42548,0.51686,1.02967,0.00078715,0.00078715,0.00078715
 45,170.612,0.34688,0.34634,0.85111,0.82016,0.80575,0.80241,0.68495,0.36745,0.48449,0.97344,0.0007822,0.0007822,0.0007822
 46,174.439,0.34067,0.33941,0.84611,0.68763,0.63768,0.55031,0.50688,0.40343,0.96323,0.9866,0.00077725,0.00077725,0.00077725
 47,178.238,0.33376,0.33747,0.84609,0.73261,0.54589,0.45909,0.42416,0.39716,0.62249,0.99632,0.0007723,0.0007723,0.0007723
 48,182.099,0.34083,0.35029,0.84964,0.45613,0.907,0.62505,0.58547,0.389,0.62485,1.02734,0.00076735,0.00076735,0.00076735
 49,185.857,0.33875,0.34073,0.84488,0.5538,0.97125,0.73572,0.67965,0.4286,0.56506,1.04781,0.0007624,0.0007624,0.0007624
 50,189.649,0.33971,0.3446,0.84582,0.46478,0.96956,0.63921,0.59861,0.37494,0.58861,1.0138,0.00075745,0.00075745,0.00075745
 51,193.425,0.33333,0.33885,0.84598,0.42493,0.88919,0.55219,0.51808,0.37723,0.54012,0.98739,0.0007525,0.0007525,0.0007525
 52,197.189,0.33127,0.32772,0.84794,0.41522,0.87923,0.53024,0.501,0.37339,0.52385,1.01347,0.00074755,0.00074755,0.00074755
 53,200.96,0.32881,0.32422,0.84444,0.78591,0.657,0.56131,0.53045,0.38148,0.61069,0.98326,0.0007426,0.0007426,0.0007426
 54,204.724,0.33491,0.33285,0.84434,0.54206,0.8907,0.60453,0.54891,0.37838,0.55873,0.94113,0.00073765,0.00073765,0.00073765
 55,208.586,0.32978,0.31757,0.8442,0.47118,0.82367,0.63179,0.59164,0.39815,0.53203,1.01263,0.0007327,0.0007327,0.0007327
 56,212.348,0.32717,0.31977,0.84713,0.45626,0.71256,0.62966,0.59155,0.36797,0.58667,0.98521,0.00072775,0.00072775,0.00072775
 57,216.11,0.33416,0.32059,0.85088,0.90537,0.54589,0.68864,0.62083,0.38039,0.61425,0.97097,0.0007228,0.0007228,0.0007228
 58,219.878,0.32603,0.30801,0.84801,0.45599,0.87923,0.60492,0.54484,0.36217,0.55401,1.01951,0.00071785,0.00071785,0.00071785
 59,223.651,0.31877,0.30675,0.84024,0.65102,0.68376,0.65738,0.57631,0.38,0.57301,1.03946,0.0007129,0.0007129,0.0007129
 60,227.443,0.31088,0.3026,0.83746,0.45508,0.76812,0.57333,0.53293,0.36359,0.49138,1.03077,0.00070795,0.00070795,0.00070795
 61,231.242,0.30442,0.30168,0.83435,0.51115,0.87923,0.58043,0.53275,0.35891,0.53742,0.96787,0.000703,0.000703,0.000703
 62,235.126,0.31241,0.31661,0.83963,0.81685,0.60756,0.60694,0.57732,0.35884,0.49992,1.00096,0.00069805,0.00069805,0.00069805
 63,238.886,0.31524,0.31796,0.8478,0.43575,0.99034,0.59844,0.56151,0.37131,0.47536,1.0316,0.0006931,0.0006931,0.0006931
 64,242.679,0.30862,0.30773,0.84656,0.40304,0.99034,0.54258,0.51098,0.35649,0.56321,0.98727,0.00068815,0.00068815,0.00068815
 65,246.438,0.30958,0.31066,0.84292,0.50694,0.67212,0.60264,0.57289,0.35521,0.52202,0.97701,0.0006832,0.0006832,0.0006832
 66,250.238,0.30157,0.28709,0.83011,0.84147,0.65274,0.64155,0.60298,0.36463,0.53978,0.97433,0.00067825,0.00067825,0.00067825
 67,254.001,0.31192,0.30306,0.84082,0.93871,0.52551,0.62054,0.58323,0.38031,0.49555,1.03516,0.0006733,0.0006733,0.0006733
 68,257.764,0.3083,0.29145,0.83946,0.41267,0.65563,0.49091,0.45388,0.35362,0.572,1.01272,0.00066835,0.00066835,0.00066835
 69,261.69,0.297,0.28773,0.83901,0.388,0.99034,0.46506,0.42749,0.36902,0.52806,0.99814,0.0006634,0.0006634,0.0006634
 70,265.486,0.30396,0.29,0.83244,0.41316,0.907,0.5592,0.52393,0.35276,0.53041,1.00164,0.00065845,0.00065845,0.00065845
 71,269.287,0.29567,0.2962,0.8369,0.48873,0.78401,0.65805,0.62952,0.33134,0.47036,0.98296,0.0006535,0.0006535,0.0006535
 72,273.053,0.28875,0.27973,0.83263,0.92451,0.65401,0.67566,0.64788,0.35246,0.46926,0.99116,0.00064855,0.00064855,0.00064855
 73,276.836,0.30926,0.29152,0.83938,0.94454,0.62615,0.67927,0.6492,0.3575,0.48188,1.0036,0.0006436,0.0006436,0.0006436
 74,280.599,0.29516,0.28729,0.82975,0.4088,0.907,0.5947,0.55653,0.37636,0.47913,1.012,0.00063865,0.00063865,0.00063865
 75,284.401,0.29825,0.28204,0.83118,0.76782,0.657,0.57365,0.53905,0.36436,0.52172,0.98047,0.0006337,0.0006337,0.0006337
 76,288.164,0.30364,0.27969,0.83378,0.74618,0.657,0.57083,0.53337,0.38702,0.56707,0.99092,0.00062875,0.00062875,0.00062875
 77,292.04,0.30469,0.28917,0.83486,0.723,0.65513,0.49255,0.45475,0.40162,0.55722,0.9436,0.0006238,0.0006238,0.0006238
 78,295.805,0.30503,0.28946,0.83452,0.42517,0.99034,0.57972,0.53013,0.37775,0.54081,0.95255,0.00061885,0.00061885,0.00061885
 79,299.604,0.27727,0.28149,0.81952,0.42256,0.86198,0.59303,0.53741,0.38325,0.47779,0.93444,0.0006139,0.0006139,0.0006139
 80,303.412,0.2982,0.28413,0.84124,0.76905,0.657,0.59481,0.56201,0.36363,0.46928,0.95052,0.00060895,0.00060895,0.00060895
 81,307.176,0.28811,0.28308,0.83208,0.43194,0.99034,0.68678,0.65044,0.35933,0.58997,0.98784,0.000604,0.000604,0.000604
 82,310.94,0.29322,0.27826,0.83489,0.90202,0.6402,0.67176,0.63767,0.38086,0.54466,1.01591,0.00059905,0.00059905,0.00059905
 83,314.739,0.29175,0.27783,0.83531,0.80618,0.62509,0.63513,0.591,0.40375,0.54441,1.03386,0.0005941,0.0005941,0.0005941
 84,318.615,0.28532,0.27561,0.82719,0.66066,0.88826,0.71195,0.57957,0.39486,0.45334,1.06155,0.00058915,0.00058915,0.00058915
 85,322.399,0.29225,0.27216,0.83676,0.93693,0.64174,0.68323,0.64803,0.38733,0.46995,1.00487,0.0005842,0.0005842,0.0005842
 86,326.164,0.29155,0.26992,0.83648,0.85519,0.65217,0.68,0.62961,0.3847,0.50255,0.99588,0.00057925,0.00057925,0.00057925
 87,329.965,0.28596,0.27408,0.83159,0.85288,0.65528,0.6306,0.59142,0.38535,0.54148,1.02609,0.0005743,0.0005743,0.0005743
 88,333.736,0.2923,0.26711,0.8351,0.62956,0.77894,0.56739,0.52419,0.40099,0.60628,1.03128,0.00056935,0.00056935,0.00056935
 89,337.504,0.28511,0.26737,0.83583,0.65626,0.67211,0.61459,0.55563,0.40065,0.57779,0.9936,0.0005644,0.0005644,0.0005644
 90,341.266,0.28832,0.27167,0.83106,0.52501,0.80435,0.66617,0.58728,0.40043,0.51282,1.00364,0.00055945,0.00055945,0.00055945
 91,345.172,0.28728,0.27036,0.84125,0.47291,0.82367,0.68532,0.63272,0.39919,0.4971,1.01576,0.0005545,0.0005545,0.0005545
 92,348.973,0.28876,0.27378,0.82925,0.42094,0.74034,0.6084,0.57226,0.41157,0.5375,1.02925,0.00054955,0.00054955,0.00054955
 93,352.744,0.29799,0.26788,0.8365,0.41912,0.99034,0.55537,0.5168,0.39969,0.50937,1.00121,0.0005446,0.0005446,0.0005446
 94,356.549,0.28235,0.26404,0.83091,0.47193,0.97585,0.61307,0.56118,0.38277,0.49662,0.98476,0.00053965,0.00053965,0.00053965
 95,360.353,0.27521,0.25805,0.81888,0.4496,0.90599,0.60918,0.56137,0.40076,0.57056,1.03051,0.0005347,0.0005347,0.0005347
--- a/fashion_yolo/finetuning_exp/finetuning_dataset.yaml
+++ b/fashion_yolo/finetuning_exp/finetuning_dataset.yaml
@ -0,0 +1,25 @@
 # path: /home/cuuva/git/Detection_Experiment/datasets/finetune_dataset/
 train: /home/cuuva/git/Detection_Experiment/datasets/finetune_dataset/train/images/
 val: /home/cuuva/git/Detection_Experiment/datasets/finetune_dataset/val/images/
 # Classes
 nc: 16  # (18개 -> 16개로 감소)
 names:
  0: shirt(blouse)
  1: t-shirt(sweatshirt)
  2: sweater
  3: cardigan
  4: jacket
  5: vest
  6: pants
  7: shorts
  8: skirt
  9: coat
  10: dress          # (dress + jumpsuit + cape)
  11: glasses        # (Old 13)
  12: hat            # (Old 14)
  13: shoe           # (Old 15)
  14: bag, wallet    # (Old 16)
  15: scarf          # (Old 17)
--- a/fashion_yolo/finetuning_exp/train.py
+++ b/fashion_yolo/finetuning_exp/train.py
@ -0,0 +1,18 @@
 from ultralytics import YOLO
 # 1. 모델 로드 (YOLOv8m 사용)
 # model = YOLO('yolov8m.pt')
 model = YOLO("/home/cuuva/git/Detection_Experiment/fashion_yolo/fashionpedia_exp/yolov8m_fashion_final/weights/best_fashion_16class.pt")
 train_results = model.train(
    data="/home/cuuva/git/Detection_Experiment/fashion_yolo/finetuning_exp/finetuning_dataset.yaml",
    epochs=200,            
    imgsz=640,
    batch=-1,              
    device="cuda",
    optimizer='AdamW',
    lr0=0.001,
    patience=50,
    project='finetuned_exp',
    name='yolov8m_finetuned_16class',
 )
--- a/fashion_yolo/init_process.ipynb
+++ b/fashion_yolo/init_process.ipynb
--- a/fashion_yolo/make_qunat.ipynb
+++ b/fashion_yolo/make_qunat.ipynb
--- a/fashion_yolo/train.py
+++ b/fashion_yolo/train.py
@ -20,7 +20,7 @@ model = YOLO('yolov8m.pt')
 # )
 train_results = model.train(
-    data="/home/cuuva/experiment/datasets/fashionpedia_yolo/fashionpedia_reduced.yaml", 
+    data="/home/cuuva/git/Detection_Experiment/datasets/fashionpedia_yolo/fashionpedia_custom.yaml", 
    epochs=500,            
    imgsz=640,
    batch=-1,              
@ -29,5 +29,5 @@ train_results = model.train(
    lr0=0.001,
    patience=50,
    project='fashionpedia_exp',
-    name='yolov8m_fashion_final',
+    name='yolov8m_fashion_4class',
 )