{ "cells": [ { "cell_type": "code", "source": [ "from google.colab import drive\n", "drive.mount('/content/drive')\n" ], "metadata": { "id": "0jig8QEwMYJu", "outputId": "466ab000-263b-4180-f829-9e62c10898df", "colab": { "base_uri": "https://localhost:8080/" } }, "execution_count": 3, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "Mounted at /content/drive\n" ] } ] }, { "cell_type": "code", "source": [ "train_dir = '/content/drive/MyDrive/Datasets/FER-2013/train'\n", "test_dir = '/content/drive/MyDrive/Datasets/FER-2013/test'\n" ], "metadata": { "id": "i-ZDzIJ9T0L8" }, "execution_count": 4, "outputs": [] }, { "cell_type": "code", "source": [ "jaffe_path = '/content/drive/MyDrive/Datasets/jaffe'" ], "metadata": { "id": "KXj5NNfzUOE_" }, "execution_count": 5, "outputs": [] }, { "cell_type": "markdown", "source": [], "metadata": { "id": "uLvct8lRTz0k" } }, { "cell_type": "code", "source": [ "import os\n" ], "metadata": { "id": "VJBbniCny0Ul" }, "execution_count": 7, "outputs": [] }, { "cell_type": "code", "source": [ "\n", "target = \"/content/drive/MyDrive/Datasets/FER-2013/train/angry\"\n", "files = os.listdir(target)\n", "supported = [f for f in files if f.lower().endswith(('.jpg', '.jpeg', '.png', '.tif', '.tiff', '.bmp'))]\n", "\n", "print(\"✅ عدد الصور بصيغ مدعومة في angry:\", len(supported))\n" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "747Kwk7lUgPe", "outputId": "7ad8e979-e02f-4da2-c0a3-2e17086427bf" }, "execution_count": 9, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "✅ عدد الصور بصيغ مدعومة في angry: 3995\n" ] } ] }, { "cell_type": "code", "source": [ "from torch.utils.data import Dataset\n", "from PIL import Image\n", "import os\n", "\n", "# ترميز المشاعر حسب رمز داخل اسم الملف\n", "expression_labels = {\n", " 'AN': 0, 'DI': 1, 'FE': 2,\n", " 'HA': 3, 'NE': 4, 'SA': 5, 'SU': 6\n", "}\n", "\n", "class JAFEDataset(Dataset):\n", " def __init__(self, root_dir, transform=None):\n", " self.root_dir = root_dir\n", " self.transform = transform\n", " self.images = []\n", " self.labels = []\n", "\n", " for file in os.listdir(root_dir):\n", " if file.lower().endswith(('.tiff', '.jpg', '.jpeg', '.png')):\n", " for code in expression_labels:\n", " if f\".{code}\" in file:\n", " self.images.append(os.path.join(root_dir, file))\n", " self.labels.append(expression_labels[code])\n", " break\n", "\n", " def __len__(self):\n", " return len(self.images)\n", "\n", " def __getitem__(self, idx):\n", " img = Image.open(self.images[idx]).convert('L')\n", " if self.transform:\n", " img = self.transform(img)\n", " return img, self.labels[idx]\n" ], "metadata": { "id": "Cdkv6by6MXT3" }, "execution_count": 10, "outputs": [] }, { "cell_type": "code", "source": [ "jaffe_transform = transforms.Compose([\n", " transforms.Resize((48, 48)),\n", " transforms.ToTensor()\n", "])\n", "\n", "jaffe_dataset = JAFEDataset(jaffe_path, transform=jaffe_transform)\n", "jaffe_loader = DataLoader(jaffe_dataset, batch_size=16, shuffle=True)\n" ], "metadata": { "id": "yr1ScWrCy5Y8" }, "execution_count": 11, "outputs": [] }, { "cell_type": "code", "source": [ "\n", "from google.colab import drive\n", "drive.mount('/content/drive')\n", "\n", "\n", "import os\n", "import torch\n", "import torchvision.transforms as transforms\n", "from torchvision.datasets import ImageFolder\n", "from torch.utils.data import DataLoader, Dataset\n", "from PIL import Image\n", "import torch.nn as nn\n", "import torchvision.models as models\n", "import torch.optim as optim\n", "from sklearn.metrics import accuracy_score\n", "\n", "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n", "\n", "\n", "transform = transforms.Compose([\n", " transforms.Grayscale(),\n", " transforms.Resize((48, 48)),\n", " transforms.ToTensor()\n", "])" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "bGM7PIqEzHOw", "outputId": "9e868125-040b-43cc-f60b-068a542d785b" }, "execution_count": 12, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n" ] } ] }, { "cell_type": "code", "source": [ "\n", "expression_labels = {\n", " 'AN': 0, 'DI': 1, 'FE': 2,\n", " 'HA': 3, 'NE': 4, 'SA': 5, 'SU': 6\n", "}\n", "\n", "class JAFEDataset(Dataset):\n", " def __init__(self, root_dir, transform=None):\n", " self.images, self.labels = [], []\n", " self.transform = transform\n", " for file in os.listdir(root_dir):\n", " if file.lower().endswith(('.jpg', '.jpeg', '.png', '.tiff')):\n", " for code in expression_labels:\n", " if f\".{code}\" in file:\n", " self.images.append(os.path.join(root_dir, file))\n", " self.labels.append(expression_labels[code])\n", " break\n", "\n", " def __len__(self):\n", " return len(self.images)\n", "\n", " def __getitem__(self, idx):\n", " img = Image.open(self.images[idx]).convert('L')\n", " if self.transform:\n", " img = self.transform(img)\n", " return img, self.labels[idx]\n", "\n", "\n", "jaffe_path = '/content/drive/MyDrive/Datasets/jaffe'\n", "jaffe_dataset = JAFEDataset(jaffe_path, transform=transform)\n", "jaffe_loader = DataLoader(jaffe_dataset, batch_size=16, shuffle=True)\n" ], "metadata": { "id": "2V6DDYUjzLl8" }, "execution_count": 13, "outputs": [] }, { "cell_type": "code", "source": [ "\n", "def train_model(train_loader, num_classes):\n", " model = models.resnet18(pretrained=True)\n", " model.conv1 = nn.Conv2d(1, 64, kernel_size=7, stride=2, padding=3, bias=False)\n", " model.fc = nn.Linear(model.fc.in_features, num_classes)\n", " model = model.to(device)\n", "\n", " optimizer = optim.Adam(model.parameters(), lr=0.0001)\n", " criterion = nn.CrossEntropyLoss()\n", "\n", " for epoch in range(3):\n", " model.train()\n", " total_loss = 0\n", " for imgs, labels in train_loader:\n", " imgs, labels = imgs.to(device), labels.to(device)\n", " optimizer.zero_grad()\n", " outputs = model(imgs)\n", " loss = criterion(outputs, labels)\n", " loss.backward()\n", " optimizer.step()\n", " total_loss += loss.item()\n", " print(f\"[Epoch {epoch+1}] Loss: {total_loss:.4f}\")\n", " return model\n", "\n", "\n", "def evaluate_model(model, test_loader):\n", " model.eval()\n", " preds, labels = [], []\n", " with torch.no_grad():\n", " for imgs, lbls in test_loader:\n", " imgs = imgs.to(device)\n", " outputs = model(imgs)\n", " _, predicted = torch.max(outputs, 1)\n", " preds.extend(predicted.cpu().numpy())\n", " labels.extend(lbls.numpy())\n", " acc = accuracy_score(labels, preds)\n", " return acc\n" ], "metadata": { "id": "-unP1GJD0E3P" }, "execution_count": 14, "outputs": [] }, { "cell_type": "code", "source": [ "import os, shutil\n", "\n", "filtered_train = '/content/filtered_train'\n", "filtered_test = '/content/filtered_test'\n", "os.makedirs(filtered_train, exist_ok=True)\n", "os.makedirs(filtered_test, exist_ok=True)\n", "\n", "valid_classes = ['sad', 'surprise']\n", "\n", "for cls in valid_classes:\n", " shutil.copytree(\n", " f'/content/drive/MyDrive/Datasets/FER-2013/train/{cls}',\n", " f'{filtered_train}/{cls}',\n", " dirs_exist_ok=True\n", " )\n", " shutil.copytree(\n", " f'/content/drive/MyDrive/Datasets/FER-2013/test/{cls}',\n", " f'{filtered_test}/{cls}',\n", " dirs_exist_ok=True\n", " )\n" ], "metadata": { "id": "xGomTYdy0RkY" }, "execution_count": 15, "outputs": [] }, { "cell_type": "code", "source": [ "from torchvision.datasets import ImageFolder\n", "from torch.utils.data import DataLoader\n", "import torchvision.transforms as transforms\n", "\n", "transform = transforms.Compose([\n", " transforms.Grayscale(),\n", " transforms.Resize((48, 48)),\n", " transforms.ToTensor()\n", "])\n", "\n", "\n", "fer_dataset_train = ImageFolder(filtered_train, transform=transform)\n", "fer_dataset_test = ImageFolder(filtered_test, transform=transform)\n", "\n", "fer_loader_train = DataLoader(fer_dataset_train, batch_size=16, shuffle=True)\n", "fer_loader_test = DataLoader(fer_dataset_test, batch_size=16, shuffle=False)\n", "\n", "print(\"✅ عدد الفئات:\", len(fer_dataset_train.classes))\n", "print(\"📦 صور تدريب:\", len(fer_dataset_train))\n", "print(\"📦 صور اختبار:\", len(fer_dataset_test))\n" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "fjoVY13Z0q6O", "outputId": "94f2dc03-f279-4847-e292-ba58679148bb" }, "execution_count": 17, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "✅ عدد الفئات: 2\n", "📦 صور تدريب: 4980\n", "📦 صور اختبار: 2078\n" ] } ] }, { "cell_type": "code", "source": [ "print(\"\\n🟦 تدريب على FER (sad + surprise فقط)\")\n", "model_fer = train_model(fer_loader_train, num_classes=len(fer_dataset_train.classes))\n", "acc_fer = evaluate_model(model_fer, fer_loader_test)\n", "print(f\"✅ دقة FER المؤقتة: {acc_fer:.2f}\")\n" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "3H8-9Kyh3KJn", "outputId": "87437510-b121-41d2-a963-54aa06e43605" }, "execution_count": 18, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "\n", "🟦 تدريب على FER (sad + surprise فقط)\n" ] }, { "output_type": "stream", "name": "stderr", "text": [ "/usr/local/lib/python3.11/dist-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.\n", " warnings.warn(\n", "/usr/local/lib/python3.11/dist-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=ResNet18_Weights.IMAGENET1K_V1`. You can also use `weights=ResNet18_Weights.DEFAULT` to get the most up-to-date weights.\n", " warnings.warn(msg)\n", "Downloading: \"https://download.pytorch.org/models/resnet18-f37072fd.pth\" to /root/.cache/torch/hub/checkpoints/resnet18-f37072fd.pth\n", "100%|██████████| 44.7M/44.7M [00:00<00:00, 137MB/s]\n" ] }, { "output_type": "stream", "name": "stdout", "text": [ "[Epoch 1] Loss: 171.5703\n", "[Epoch 2] Loss: 120.5325\n", "[Epoch 3] Loss: 89.5785\n", "✅ دقة FER المؤقتة: 0.83\n" ] } ] }, { "cell_type": "code", "source": [ "model_jaffe = train_model(jaffe_loader, num_classes=7)\n", "acc_jaffe = evaluate_model(model_jaffe, jaffe_loader)\n", "print(f\"✅ دقة JAFFE: {acc_jaffe:.2f}\")\n" ], "metadata": { "id": "Xu5C8jX41Hrg", "colab": { "base_uri": "https://localhost:8080/" }, "outputId": "6c605b2b-7af2-4323-8180-a25e7387a08a" }, "execution_count": 19, "outputs": [ { "output_type": "stream", "name": "stderr", "text": [ "/usr/local/lib/python3.11/dist-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.\n", " warnings.warn(\n", "/usr/local/lib/python3.11/dist-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=ResNet18_Weights.IMAGENET1K_V1`. You can also use `weights=ResNet18_Weights.DEFAULT` to get the most up-to-date weights.\n", " warnings.warn(msg)\n" ] }, { "output_type": "stream", "name": "stdout", "text": [ "[Epoch 1] Loss: 30.2210\n", "[Epoch 2] Loss: 19.0657\n", "[Epoch 3] Loss: 11.7811\n", "✅ دقة JAFFE: 0.17\n" ] } ] }, { "cell_type": "code", "source": [ "print(\"📊 مقارنة النتائج:\")\n", "print(f\"✔️ FER (sad + surprise): {acc_fer:.2f}\")\n", "print(f\"✔️ JAFFE: {acc_jaffe:.2f}\")\n" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "aCFtn36Z4hE1", "outputId": "ff82b439-e94f-4be9-f1ec-4a75981b76b8" }, "execution_count": 21, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "📊 مقارنة النتائج:\n", "✔️ FER (sad + surprise): 0.83\n", "✔️ JAFFE: 0.17\n" ] } ] }, { "cell_type": "code", "source": [ "import pandas as pd\n", "\n", "csv_path = '/content/drive/MyDrive/Datasets/CK+/ckextended.csv'\n", "\n", "df = pd.read_csv(csv_path)\n", "print(df.head())\n" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "oV7xbFVU48Km", "outputId": "3776318a-23cb-43ac-fe0e-6e8d18138c1d" }, "execution_count": 22, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ " emotion pixels Usage\n", "0 6 36 39 35 25 19 11 8 7 3 13 15 9 21 57 75 90 10... Training\n", "1 6 88 74 19 4 5 5 3 12 8 21 15 21 15 18 24 29 32 ... Training\n", "2 6 9 2 4 7 1 1 1 0 7 29 49 76 115 141 156 169 177... Training\n", "3 6 104 106 108 104 95 50 60 61 58 83 126 133 139 ... Training\n", "4 6 68 72 67 67 6 2 1 1 1 1 1 14 24 24 38 65 79 94... Training\n" ] } ] }, { "cell_type": "code", "source": [ "import numpy as np\n", "from torch.utils.data import Dataset\n", "from PIL import Image\n", "\n", "class CKDatasetFromCSV(Dataset):\n", " def __init__(self, csv_path, usage=\"Training\", transform=None):\n", " self.data = pd.read_csv(csv_path)\n", " self.data = self.data[self.data['Usage'] == usage]\n", " self.transform = transform\n", "\n", " def __len__(self):\n", " return len(self.data)\n", "\n", " def __getitem__(self, idx):\n", " pixels = list(map(int, self.data.iloc[idx]['pixels'].split()))\n", " img = np.array(pixels).reshape(48, 48).astype('uint8')\n", " img = Image.fromarray(img)\n", " label = int(self.data.iloc[idx]['emotion'])\n", "\n", " if self.transform:\n", " img = self.transform(img)\n", "\n", " return img, label\n" ], "metadata": { "id": "ryMksQR-5Eyi" }, "execution_count": 23, "outputs": [] }, { "cell_type": "code", "source": [ "transform = transforms.Compose([\n", " transforms.Resize((48, 48)),\n", " transforms.ToTensor()\n", "])\n", "\n", "ck_csv_path = '/content/drive/MyDrive/Datasets/CK+/ckextended.csv'\n", "\n", "ck_dataset = CKDatasetFromCSV(csv_path=ck_csv_path, usage=\"Training\", transform=transform)\n", "ck_loader = DataLoader(ck_dataset, batch_size=16, shuffle=True)\n", "\n", "print(\"✅ صور CK+: \", len(ck_dataset))\n" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "G6cxN7_S5VGb", "outputId": "e214a07a-979e-454d-d2d7-727cc71cf69b" }, "execution_count": 24, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "✅ صور CK+: 734\n" ] } ] }, { "cell_type": "code", "source": [ "df = pd.read_csv('/content/drive/MyDrive/Datasets/CK+/ckextended.csv')\n", "print(\"الفئات الموجودة:\", sorted(df['emotion'].unique()))\n" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "f0bsjaJs5YCr", "outputId": "14d36574-acca-41da-f35d-ada3386f15d7" }, "execution_count": 25, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "الفئات الموجودة: [np.int64(0), np.int64(1), np.int64(2), np.int64(3), np.int64(4), np.int64(5), np.int64(6), np.int64(7)]\n" ] } ] }, { "cell_type": "code", "source": [ "model_ck = train_model(ck_loader, num_classes=8)\n" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "eaQDVX3R7Afl", "outputId": "145812b5-cf69-44e5-c877-b50a6f8043ad" }, "execution_count": 26, "outputs": [ { "output_type": "stream", "name": "stderr", "text": [ "/usr/local/lib/python3.11/dist-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and may be removed in the future, please use 'weights' instead.\n", " warnings.warn(\n", "/usr/local/lib/python3.11/dist-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or `None` for 'weights' are deprecated since 0.13 and may be removed in the future. The current behavior is equivalent to passing `weights=ResNet18_Weights.IMAGENET1K_V1`. You can also use `weights=ResNet18_Weights.DEFAULT` to get the most up-to-date weights.\n", " warnings.warn(msg)\n" ] }, { "output_type": "stream", "name": "stdout", "text": [ "[Epoch 1] Loss: 76.9791\n", "[Epoch 2] Loss: 29.1143\n", "[Epoch 3] Loss: 16.6405\n" ] } ] }, { "cell_type": "code", "source": [ "ck_dataset = CKDatasetFromCSV(csv_path=ck_csv_path, usage=\"Training\", transform=transform)\n" ], "metadata": { "id": "g2s0j96M7bFD" }, "execution_count": 27, "outputs": [] }, { "cell_type": "code", "source": [ "jaffe_dataset = JAFEDataset(jaffe_path, transform=transform)\n" ], "metadata": { "id": "2pq_oa9O7TzB" }, "execution_count": 28, "outputs": [] }, { "cell_type": "code", "source": [ "fer_train_dir = '/content/drive/MyDrive/Datasets/FER-2013/train'\n", "fer_test_dir = '/content/drive/MyDrive/Datasets/FER-2013/test'\n", "\n", "jaffe_dir = '/content/drive/MyDrive/Datasets/jaffe'\n", "ck_csv_path = '/content/drive/MyDrive/Datasets/CK+/ckextended.csv'\n" ], "metadata": { "id": "xQbMgRI79V8P" }, "execution_count": 29, "outputs": [] }, { "cell_type": "code", "source": [ "import os\n", "import numpy as np\n", "import pandas as pd\n", "from PIL import Image\n", "import matplotlib.pyplot as plt\n", "\n", "import torch\n", "import torch.nn as nn\n", "import torch.optim as optim\n", "from torch.utils.data import Dataset, DataLoader, ConcatDataset\n", "import torchvision.transforms as transforms\n", "from torchvision.datasets import ImageFolder\n", "import torchvision.models as models\n" ], "metadata": { "id": "YOn-HStz9XYZ" }, "execution_count": 31, "outputs": [] }, { "cell_type": "code", "source": [ "\n", "transform = transforms.Compose([\n", " transforms.Grayscale(num_output_channels=1),\n", " transforms.Resize((48, 48)),\n", " transforms.ToTensor()\n", "])\n" ], "metadata": { "id": "Ya1Pmt869w-6" }, "execution_count": 32, "outputs": [] }, { "cell_type": "code", "source": [ "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n", "print(\"✅ Running on:\", device)\n" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "-KcjQ6VF_7oa", "outputId": "5ede571e-75f8-4eac-abc8-aa540b8a8813" }, "execution_count": 33, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "✅ Running on: cpu\n" ] } ] }, { "cell_type": "code", "source": [ "target = \"/content/drive/MyDrive/Datasets/FER-2013/train/angry\"\n", "files = os.listdir(target)\n", "supported = [f for f in files if f.lower().endswith(('.jpg', '.jpeg', '.png', '.tif', '.tiff', '.bmp'))]\n", "print(f\"📦 عدد الصور المدعومة في angry: {len(supported)}\")\n" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "HJ0gz1Rb_-wI", "outputId": "bdb30da1-b3d4-4971-b1e2-8200cb98aeb1" }, "execution_count": 34, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "📦 عدد الصور المدعومة في angry: 3995\n" ] } ] }, { "cell_type": "code", "source": [ "import torch.nn as nn\n", "import torch.nn.functional as F\n", "\n", "class DepthwiseSeparableConv(nn.Module):\n", " def __init__(self, in_channels, out_channels):\n", " super().__init__()\n", " self.depthwise = nn.Conv2d(in_channels, in_channels, kernel_size=3, padding=1, groups=in_channels, bias=False)\n", " self.pointwise = nn.Conv2d(in_channels, out_channels, kernel_size=1, bias=False)\n", " self.bn = nn.BatchNorm2d(out_channels)\n", "\n", " def forward(self, x):\n", " x = self.depthwise(x)\n", " x = self.pointwise(x)\n", " x = self.bn(x)\n", " return F.relu(x)\n", "\n", "class MiniXception(nn.Module):\n", " def __init__(self, num_classes=7):\n", " super().__init__()\n", " self.entry = nn.Sequential(\n", " nn.Conv2d(1, 32, kernel_size=3, stride=1, padding=1),\n", " nn.BatchNorm2d(32),\n", " nn.ReLU(),\n", " nn.Conv2d(32, 64, kernel_size=3, stride=1, padding=1),\n", " nn.BatchNorm2d(64),\n", " nn.ReLU()\n", " )\n", "\n", " self.module1 = DepthwiseSeparableConv(64, 128)\n", " self.module2 = DepthwiseSeparableConv(128, 256)\n", " self.module3 = DepthwiseSeparableConv(256, 512)\n", " self.pool = nn.AdaptiveAvgPool2d((1, 1))\n", " self.fc = nn.Linear(512, num_classes)\n", "\n", " def forward(self, x):\n", " x = self.entry(x)\n", " x = self.module1(x)\n", " x = self.module2(x)\n", " x = self.module3(x)\n", " x = self.pool(x)\n", " x = x.view(x.size(0), -1)\n", " return self.fc(x)\n" ], "metadata": { "id": "C6oGWXHBADHn" }, "execution_count": 35, "outputs": [] }, { "cell_type": "code", "source": [ "def train_model_minix(model, loader, num_epochs=5, lr=0.001):\n", " model = model.to(device)\n", " optimizer = optim.Adam(model.parameters(), lr=lr)\n", " criterion = nn.CrossEntropyLoss()\n", "\n", " for epoch in range(num_epochs):\n", " model.train()\n", " total_loss = 0\n", " for imgs, labels in loader:\n", " imgs, labels = imgs.to(device), labels.to(device)\n", " optimizer.zero_grad()\n", " outputs = model(imgs)\n", " loss = criterion(outputs, labels)\n", " loss.backward()\n", " optimizer.step()\n", " total_loss += loss.item()\n", " print(f\"[Epoch {epoch+1}] Loss: {total_loss:.4f}\")\n", " return model\n" ], "metadata": { "id": "rM1JMEuhAp-f" }, "execution_count": 36, "outputs": [] }, { "cell_type": "code", "source": [ "from sklearn.metrics import accuracy_score\n", "\n", "def evaluate_model(model, loader):\n", " model.eval()\n", " preds, true = [], []\n", " with torch.no_grad():\n", " for imgs, labels in loader:\n", " imgs = imgs.to(device)\n", " outputs = model(imgs)\n", " _, predicted = torch.max(outputs, 1)\n", " preds.extend(predicted.cpu().numpy())\n", " true.extend(labels.numpy())\n", " acc = accuracy_score(true, preds)\n", " return acc\n" ], "metadata": { "id": "GgEi4ancAuXW" }, "execution_count": 37, "outputs": [] }, { "cell_type": "code", "source": [ "model = MiniXception(num_classes=7)\n", "model = train_model_minix(model, jaffe_loader, num_epochs=5)\n", "acc = evaluate_model(model, jaffe_loader)\n", "print(f\"✅ دقة Mini-Xception على JAFFE: {acc:.2f}\")\n" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "fbTtDGGhAxMt", "outputId": "6ea1478c-d14d-4d9b-a77e-016b8d860ec1" }, "execution_count": 38, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "[Epoch 1] Loss: 28.3883\n", "[Epoch 2] Loss: 27.7037\n", "[Epoch 3] Loss: 26.6824\n", "[Epoch 4] Loss: 26.2634\n", "[Epoch 5] Loss: 25.7474\n", "✅ دقة Mini-Xception على JAFFE: 0.23\n" ] } ] }, { "cell_type": "code", "source": [ "class TransferMiniXception(nn.Module):\n", " def __init__(self, num_classes=7):\n", " super().__init__()\n", " self.features = nn.Sequential(\n", " nn.Conv2d(1, 32, kernel_size=3, stride=1, padding=1),\n", " nn.BatchNorm2d(32),\n", " nn.ReLU(),\n", " nn.Conv2d(32, 64, kernel_size=3, stride=1, padding=1),\n", " nn.BatchNorm2d(64),\n", " nn.ReLU(),\n", " DepthwiseSeparableConv(64, 128),\n", " DepthwiseSeparableConv(128, 256),\n", " DepthwiseSeparableConv(256, 512)\n", " )\n", "\n", " self.pool = nn.AdaptiveAvgPool2d((1, 1))\n", " self.classifier = nn.Linear(512, num_classes)\n", "\n", " def forward(self, x):\n", " x = self.features(x)\n", " x = self.pool(x)\n", " x = x.view(x.size(0), -1)\n", " return self.classifier(x)\n" ], "metadata": { "id": "wUd1-2svAz0f" }, "execution_count": 39, "outputs": [] }, { "cell_type": "code", "source": [ "model = TransferMiniXception(num_classes=7)\n", "\n", "for param in model.features.parameters():\n", " param.requires_grad = False\n", "\n", "model = model.to(device)\n" ], "metadata": { "id": "LvJiJDq_BoXA" }, "execution_count": 41, "outputs": [] }, { "cell_type": "code", "source": [ "def train_transfer(model, loader, num_epochs=5, lr=0.001):\n", " optimizer = optim.Adam(filter(lambda p: p.requires_grad, model.parameters()), lr=lr)\n", " criterion = nn.CrossEntropyLoss()\n", "\n", " for epoch in range(num_epochs):\n", " model.train()\n", " total_loss = 0\n", " for imgs, labels in loader:\n", " imgs, labels = imgs.to(device), labels.to(device)\n", " optimizer.zero_grad()\n", " outputs = model(imgs)\n", " loss = criterion(outputs, labels)\n", " loss.backward()\n", " optimizer.step()\n", " total_loss += loss.item()\n", " print(f\"[Epoch {epoch+1}] Loss: {total_loss:.4f}\")\n", " return model\n" ], "metadata": { "id": "qo3KsoCqBsBN" }, "execution_count": 42, "outputs": [] }, { "cell_type": "code", "source": [ "acc = evaluate_model(model, jaffe_loader)\n", "print(f\"✅ دقة Transfer Learning Mini-Xception: {acc:.2f}\")\n" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "91y2yNU8Bx05", "outputId": "2e778e81-9009-46f3-fe07-628726434feb" }, "execution_count": 43, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "✅ دقة Transfer Learning Mini-Xception: 0.14\n" ] } ] }, { "cell_type": "code", "source": [ "from torchvision.models import resnet18, ResNet18_Weights\n", "\n", "model = resnet18(weights=ResNet18_Weights.DEFAULT)\n", "\n", "for param in model.parameters():\n", " param.requires_grad = False\n" ], "metadata": { "id": "-sZpKzgqB0-4" }, "execution_count": 44, "outputs": [] }, { "cell_type": "code", "source": [ "num_classes = 7\n", "\n", "in_features = model.fc.in_features\n", "model.fc = nn.Linear(in_features, num_classes)\n", "\n", "model = model.to(device)\n" ], "metadata": { "id": "KTWTzY_BCENN" }, "execution_count": 46, "outputs": [] }, { "cell_type": "code", "source": [ "def train_resnet(model, loader, num_epochs=5, lr=0.001):\n", " criterion = nn.CrossEntropyLoss()\n", " optimizer = optim.Adam(model.fc.parameters(), lr=lr) # فقط طبقة fc\n", "\n", " for epoch in range(num_epochs):\n", " model.train()\n", " total_loss = 0\n", " for imgs, labels in loader:\n", " imgs, labels = imgs.to(device), labels.to(device)\n", " optimizer.zero_grad()\n", " outputs = model(imgs)\n", " loss = criterion(outputs, labels)\n", " loss.backward()\n", " optimizer.step()\n", " total_loss += loss.item()\n", " print(f\"[Epoch {epoch+1}] Loss: {total_loss:.4f}\")\n", " return model\n" ], "metadata": { "id": "XNNDnAEDCH4b" }, "execution_count": 47, "outputs": [] }, { "cell_type": "code", "source": [ "transform = transforms.Compose([\n", " transforms.Resize((48, 48)),\n", " transforms.Grayscale(num_output_channels=3),\n", " transforms.ToTensor(),\n", "])\n" ], "metadata": { "id": "ZZaml9fICMOc" }, "execution_count": 49, "outputs": [] }, { "cell_type": "code", "source": [ "jaffe_dataset = JAFEDataset(jaffe_path, transform=transform)\n", "jaffe_loader = DataLoader(jaffe_dataset, batch_size=16, shuffle=True)\n" ], "metadata": { "id": "VdUnaRnOCPTw" }, "execution_count": 50, "outputs": [] }, { "cell_type": "code", "source": [ "model = resnet18(weights=ResNet18_Weights.DEFAULT)\n", "for param in model.parameters():\n", " param.requires_grad = False\n", "\n", "model.fc = nn.Linear(model.fc.in_features, 7)\n", "model = model.to(device)\n", "\n", "model = train_resnet(model, jaffe_loader, num_epochs=5)\n", "acc = evaluate_model(model, jaffe_loader)\n", "print(f\"✅ دقة ResNet18 على JAFFE: {acc:.2f}\")\n" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "OLt1dKIECg6d", "outputId": "1461adf8-3f64-463b-f2d7-e94efd8e885d" }, "execution_count": 51, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "[Epoch 1] Loss: 28.6744\n", "[Epoch 2] Loss: 25.0066\n", "[Epoch 3] Loss: 22.2953\n", "[Epoch 4] Loss: 20.3337\n", "[Epoch 5] Loss: 19.7773\n", "✅ دقة ResNet18 على JAFFE: 0.41\n" ] } ] }, { "cell_type": "code", "source": [ "transform = transforms.Compose([\n", " transforms.Resize((48, 48)),\n", " transforms.Grayscale(num_output_channels=3), # ← يحول الصور من 1 إلى 3 قنوات\n", " transforms.ToTensor(),\n", "])\n" ], "metadata": { "id": "QFrv8ALNCitl" }, "execution_count": 52, "outputs": [] }, { "cell_type": "code", "source": [ "jaffe_dataset = JAFEDataset(jaffe_path, transform=transform)\n", "jaffe_loader = DataLoader(jaffe_dataset, batch_size=16, shuffle=True)" ], "metadata": { "id": "4WfEMfB6EyUW" }, "execution_count": 53, "outputs": [] }, { "cell_type": "code", "source": [ "def __getitem__(self, idx):\n", " row = self.data.iloc[idx]\n", "\n", " # تحويل سلسلة البكسلات إلى صورة NumPy 48x48\n", " pixels = np.fromstring(row['pixels'], dtype=int, sep=' ')\n", " img_array = pixels.reshape(48, 48) # ← حجم FER وCK+ المعتاد\n", "\n", " # تحويل إلى صورة PIL\n", " img = Image.fromarray(img_array.astype(np.uint8)).convert(\"L\")\n", " img = img.convert(\"RGB\") # ← حتى يعمل ResNet18 (3 قنوات)\n", "\n", " if self.transform:\n", " img = self.transform(img)\n", "\n", " label = int(row['emotion'])\n", " return img, label\n" ], "metadata": { "id": "0WF3g509E1AA" }, "execution_count": 56, "outputs": [] }, { "cell_type": "code", "source": [ "fer_path = \"/content/drive/MyDrive/Datasets/FER-2013/train\" # أو حسب موقعك الفعلي\n" ], "metadata": { "id": "ilp1rZdEFfX0" }, "execution_count": 59, "outputs": [] }, { "cell_type": "code", "source": [ "import os\n", "\n", "fer_path = \"/content/drive/MyDrive/Datasets/FER-2013/train\"\n", "\n", "for label in os.listdir(fer_path):\n", " folder = os.path.join(fer_path, label)\n", " if os.path.isdir(folder):\n", " images = [f for f in os.listdir(folder) if f.lower().endswith(('.jpg', '.jpeg', '.png'))]\n", " print(f\"{label}: {len(images)} صورة مدعومة\")\n", "\n" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "id": "-eF5q9CvFj-G", "outputId": "8040c22e-5a60-4fed-b2a7-e23e2e23cbd7" }, "execution_count": 61, "outputs": [ { "output_type": "stream", "name": "stdout", "text": [ "surprise: 3171 صورة مدعومة\n", "happy: 0 صورة مدعومة\n", "fear: 2277 صورة مدعومة\n", "sad: 1809 صورة مدعومة\n", "neutral: 0 صورة مدعومة\n", "angry: 3995 صورة مدعومة\n", "disgust: 436 صورة مدعومة\n" ] } ] }, { "cell_type": "code", "source": [ "from torchvision import transforms\n", "\n", "transform = transforms.Compose([\n", " transforms.Resize((48, 48)),\n", " transforms.RandomHorizontalFlip(p=0.5),\n", " transforms.RandomRotation(10),\n", " transforms.RandomCrop(48, padding=4),\n", " transforms.Grayscale(num_output_channels=3),\n", " transforms.ToTensor(),\n", " transforms.Normalize([0.5]*3, [0.5]*3)\n", "])\n" ], "metadata": { "id": "jENwf-NsGY62" }, "execution_count": 62, "outputs": [] }, { "cell_type": "code", "source": [ "from torch.utils.data import DataLoader\n", "from torchvision.datasets import ImageFolder\n", "\n", "train_path = \"/content/drive/MyDrive/Datasets/FER-2013/train\" # غيّره حسب مسارك\n", "\n", "transform = transforms.Compose([\n", " transforms.Resize((48, 48)),\n", " transforms.Grayscale(num_output_channels=3),\n", " transforms.RandomHorizontalFlip(),\n", " transforms.ToTensor(),\n", " transforms.Normalize([0.5]*3, [0.5]*3)\n", "])\n", "\n", "train_dataset = ImageFolder(train_path, transform=transform)\n", "train_loader = DataLoader(train_dataset, batch_size=16, shuffle=True)\n" ], "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 391 }, "id": "3X3qxGXFHI16", "outputId": "f756691d-b3a4-4bb6-d2aa-20fd8b0c62ef" }, "execution_count": 67, "outputs": [ { "output_type": "error", "ename": "FileNotFoundError", "evalue": "Found no valid file for the classes happy, neutral. Supported extensions are: .jpg, .jpeg, .png, .ppm, .bmp, .pgm, .tif, .tiff, .webp", "traceback": [ "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m", "\u001b[0;31mFileNotFoundError\u001b[0m Traceback (most recent call last)", "\u001b[0;32m/tmp/ipython-input-67-2083617579.py\u001b[0m in \u001b[0;36m\u001b[0;34m()\u001b[0m\n\u001b[1;32m 12\u001b[0m ])\n\u001b[1;32m 13\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 14\u001b[0;31m \u001b[0mtrain_dataset\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mImageFolder\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtrain_path\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtransform\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mtransform\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 15\u001b[0m \u001b[0mtrain_loader\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mDataLoader\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtrain_dataset\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbatch_size\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m16\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mshuffle\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n", "\u001b[0;32m/usr/local/lib/python3.11/dist-packages/torchvision/datasets/folder.py\u001b[0m in \u001b[0;36m__init__\u001b[0;34m(self, root, transform, target_transform, loader, is_valid_file, allow_empty)\u001b[0m\n\u001b[1;32m 326\u001b[0m \u001b[0mallow_empty\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mbool\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 327\u001b[0m ):\n\u001b[0;32m--> 328\u001b[0;31m super().__init__(\n\u001b[0m\u001b[1;32m 329\u001b[0m \u001b[0mroot\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 330\u001b[0m \u001b[0mloader\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n", "\u001b[0;32m/usr/local/lib/python3.11/dist-packages/torchvision/datasets/folder.py\u001b[0m in \u001b[0;36m__init__\u001b[0;34m(self, root, loader, extensions, transform, target_transform, is_valid_file, allow_empty)\u001b[0m\n\u001b[1;32m 148\u001b[0m \u001b[0msuper\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__init__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mroot\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtransform\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mtransform\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtarget_transform\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mtarget_transform\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 149\u001b[0m \u001b[0mclasses\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclass_to_idx\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfind_classes\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mroot\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 150\u001b[0;31m samples = self.make_dataset(\n\u001b[0m\u001b[1;32m 151\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mroot\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 152\u001b[0m \u001b[0mclass_to_idx\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mclass_to_idx\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n", "\u001b[0;32m/usr/local/lib/python3.11/dist-packages/torchvision/datasets/folder.py\u001b[0m in \u001b[0;36mmake_dataset\u001b[0;34m(directory, class_to_idx, extensions, is_valid_file, allow_empty)\u001b[0m\n\u001b[1;32m 201\u001b[0m \u001b[0;31m# is potentially overridden and thus could have a different logic.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 202\u001b[0m \u001b[0;32mraise\u001b[0m \u001b[0mValueError\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"The class_to_idx parameter cannot be None.\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 203\u001b[0;31m return make_dataset(\n\u001b[0m\u001b[1;32m 204\u001b[0m \u001b[0mdirectory\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mclass_to_idx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mextensions\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mextensions\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mis_valid_file\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mis_valid_file\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mallow_empty\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mallow_empty\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 205\u001b[0m )\n", "\u001b[0;32m/usr/local/lib/python3.11/dist-packages/torchvision/datasets/folder.py\u001b[0m in \u001b[0;36mmake_dataset\u001b[0;34m(directory, class_to_idx, extensions, is_valid_file, allow_empty)\u001b[0m\n\u001b[1;32m 102\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mextensions\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 103\u001b[0m \u001b[0mmsg\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0;34mf\"Supported extensions are: {extensions if isinstance(extensions, str) else ', '.join(extensions)}\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 104\u001b[0;31m \u001b[0;32mraise\u001b[0m \u001b[0mFileNotFoundError\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmsg\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 105\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 106\u001b[0m \u001b[0;32mreturn\u001b[0m \u001b[0minstances\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n", "\u001b[0;31mFileNotFoundError\u001b[0m: Found no valid file for the classes happy, neutral. Supported extensions are: .jpg, .jpeg, .png, .ppm, .bmp, .pgm, .tif, .tiff, .webp" ] } ] }, { "cell_type": "code", "source": [], "metadata": { "id": "PArhXPRBHLZ1" }, "execution_count": null, "outputs": [] } ], "metadata": { "colab": { "provenance": [] }, "kernelspec": { "display_name": "Python 3", "name": "python3" } }, "nbformat": 4, "nbformat_minor": 0 }