channel-pruning-of-effficientnet-v2-s-cifar10.ipynb

{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.10.10","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"pip install torch-pruning ","metadata":{"execution":{"iopub.status.busy":"2023-07-24T09:19:45.848193Z","iopub.execute_input":"2023-07-24T09:19:45.848863Z","iopub.status.idle":"2023-07-24T09:19:58.655165Z","shell.execute_reply.started":"2023-07-24T09:19:45.848826Z","shell.execute_reply":"2023-07-24T09:19:58.653923Z"},"trusted":true},"execution_count":1,"outputs":[{"name":"stdout","text":"Collecting torch-pruning\n  Downloading torch_pruning-1.2.0-py3-none-any.whl (44 kB)\n\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m44.1/44.1 kB\u001b[0m \u001b[31m3.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n\u001b[?25hRequirement already satisfied: torch in /opt/conda/lib/python3.10/site-packages (from torch-pruning) (2.0.0)\nRequirement already satisfied: numpy in /opt/conda/lib/python3.10/site-packages (from torch-pruning) (1.23.5)\nRequirement already satisfied: filelock in /opt/conda/lib/python3.10/site-packages (from torch->torch-pruning) (3.12.0)\nRequirement already satisfied: typing-extensions in /opt/conda/lib/python3.10/site-packages (from torch->torch-pruning) (4.5.0)\nRequirement already satisfied: sympy in /opt/conda/lib/python3.10/site-packages (from torch->torch-pruning) (1.12)\nRequirement already satisfied: networkx in /opt/conda/lib/python3.10/site-packages (from torch->torch-pruning) (3.1)\nRequirement already satisfied: jinja2 in /opt/conda/lib/python3.10/site-packages (from torch->torch-pruning) (3.1.2)\nRequirement already satisfied: MarkupSafe>=2.0 in /opt/conda/lib/python3.10/site-packages (from jinja2->torch->torch-pruning) (2.1.2)\nRequirement already satisfied: mpmath>=0.19 in /opt/conda/lib/python3.10/site-packages (from sympy->torch->torch-pruning) (1.3.0)\nInstalling collected packages: torch-pruning\nSuccessfully installed torch-pruning-1.2.0\n\u001b[33mWARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv\u001b[0m\u001b[33m\n\u001b[0mNote: you may need to restart the kernel to use updated packages.\n","output_type":"stream"}]},{"cell_type":"markdown","source":"# Method\nI will be using maginutde based pruning, the group of channel with lowes L2 norm will be pruned iteratively.\nLayer-adaptive sparsity for the Magnitude-based Pruning: https://arxiv.org/abs/2010.07611\n\n","metadata":{}},{"cell_type":"code","source":"import torch \nimport torch.nn as nn\nimport torchvision\nfrom torchvision import transforms\nfrom tqdm import tqdm\nfrom torch.optim import lr_scheduler\nimport torch.optim as optim\nimport time\nimport os\nimport random\nimport shutil\n# import torch_pruning as tp","metadata":{"execution":{"iopub.status.busy":"2023-07-25T01:42:49.264342Z","iopub.execute_input":"2023-07-25T01:42:49.264710Z","iopub.status.idle":"2023-07-25T01:42:49.271044Z","shell.execute_reply.started":"2023-07-25T01:42:49.264681Z","shell.execute_reply":"2023-07-25T01:42:49.269762Z"},"trusted":true},"execution_count":20,"outputs":[]},{"cell_type":"code","source":"!pip install torchsummary","metadata":{"execution":{"iopub.status.busy":"2023-07-24T09:19:58.682863Z","iopub.execute_input":"2023-07-24T09:19:58.683193Z","iopub.status.idle":"2023-07-24T09:20:10.104754Z","shell.execute_reply.started":"2023-07-24T09:19:58.683163Z","shell.execute_reply":"2023-07-24T09:20:10.103542Z"},"trusted":true},"execution_count":3,"outputs":[{"name":"stdout","text":"Collecting torchsummary\n  Downloading torchsummary-1.5.1-py3-none-any.whl (2.8 kB)\nInstalling collected packages: torchsummary\nSuccessfully installed torchsummary-1.5.1\n\u001b[33mWARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv\u001b[0m\u001b[33m\n\u001b[0m","output_type":"stream"}]},{"cell_type":"markdown","source":"paper: https://arxiv.org/pdf/2305.14403.pdf","metadata":{}},{"cell_type":"markdown","source":"# Model Loading, Dataset Prep","metadata":{}},{"cell_type":"code","source":"model = torch.load('/kaggle/input/pruning-effficientnet-v2-s-on-cifar10/81%_pruned_unmasked_model.pt')","metadata":{"execution":{"iopub.status.busy":"2023-07-24T09:20:10.108232Z","iopub.execute_input":"2023-07-24T09:20:10.108635Z","iopub.status.idle":"2023-07-24T09:20:14.310748Z","shell.execute_reply.started":"2023-07-24T09:20:10.108592Z","shell.execute_reply":"2023-07-24T09:20:14.309778Z"},"trusted":true},"execution_count":4,"outputs":[]},{"cell_type":"code","source":"#All the necessary code to load and configure the dataset \n\nbatch_size = 32\npath = '/kaggle/input/cifar10-python'\ndevice = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\n\ndata_transforms = {\n    'train': transforms.Compose([\n        transforms.Resize(224),\n        transforms.RandomHorizontalFlip(),\n        transforms.ToTensor(),\n        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])\n    ]),\n    'test': transforms.Compose([\n        transforms.Resize(224),\n        transforms.ToTensor(),\n        transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])\n    ]),\n}\n\n\ntrainset = torchvision.datasets.CIFAR10(root=path, \n                                        train=True, download=False, \n                                        transform=data_transforms['train'])\n\ntrainloader = torch.utils.data.DataLoader(trainset, batch_size=batch_size, shuffle=True)\n\n\ntestset = torchvision.datasets.CIFAR10(root=path, \n                                        train=False, download=False, \n                                        transform=data_transforms['test'])\n\n                                     \ntestloader = torch.utils.data.DataLoader(testset, batch_size=batch_size,\n                                         shuffle=False)\n\nclasses = ('plane', 'car', 'bird', 'cat',\n           'deer', 'dog', 'frog', 'horse', 'ship', 'truck')\n\n\n\n#Making things easy, so storing it in a dictionary \ndataset_size = [len(trainset), len(testset)]\ndataloaders= {\n    \"train\": trainloader,\n    \"val\": testloader\n}\ndataset_sizes = {\n    \"train\": dataset_size[0],\n    \"val\": dataset_size[1]\n}\n\n\n\n#Evaluating the model performance on the test set\ndef evaluate_model(model, criterion = nn.CrossEntropyLoss() , dataloader = testloader ):\n    model.eval()\n    running_loss = 0.0\n    running_corrects = 0\n\n    for inputs, labels in tqdm(dataloader):\n        inputs = inputs.to(device)\n        labels = labels.to(device)\n\n        with torch.no_grad():\n            outputs = model(inputs)\n            _, preds = torch.max(outputs, 1)\n            loss = criterion(outputs, labels)\n\n        running_loss += loss.item() * inputs.size(0)\n        running_corrects += torch.sum(preds == labels.data)\n\n    eval_loss = running_loss / len(dataloader.dataset)\n    eval_acc = running_corrects.double() / len(dataloader.dataset)\n\n    print(f'Evaluation Loss: {eval_loss:.4f} Acc: {eval_acc:.4f}')\n\n    return eval_loss, eval_acc\n\n\n\n\nclass_names = trainset.classes","metadata":{"execution":{"iopub.status.busy":"2023-07-25T01:42:54.593031Z","iopub.execute_input":"2023-07-25T01:42:54.593627Z","iopub.status.idle":"2023-07-25T01:42:57.928847Z","shell.execute_reply.started":"2023-07-25T01:42:54.593587Z","shell.execute_reply":"2023-07-25T01:42:57.927790Z"},"trusted":true},"execution_count":21,"outputs":[]},{"cell_type":"markdown","source":"# Train defination ","metadata":{}},{"cell_type":"code","source":"#train & evaluate\n\n#Function for training the model, forware prop and backward prop \ndef train_model(model, criterion, optimizer, scheduler, num_epochs=25):\n    since = time.time()\n    \n    #storing epoch data\n    epoch_data =     {\n        'epoch': [],\n        'train': {'loss': [], 'acc': []},\n        'val': {'loss': [], 'acc': [] }\n    }\n    \n    # Create a temporary directory in Kaggle's temp directory\n    tempdir = '/kaggle/working/temp'\n    os.makedirs(tempdir, exist_ok=True)\n    best_model_params_path = os.path.join(tempdir, 'best_model_params.pt')\n\n    torch.save(model.state_dict(), best_model_params_path)\n    best_acc = 0.0\n\n    for epoch in range(num_epochs):\n        print(f'Epoch {epoch+1}/{num_epochs}')\n        print('-' * 10)\n        epoch_data['epoch'].append(epoch+1)\n        \n        for phase in ['train', 'val']:\n            if phase == 'train':\n                model.train()\n            else:\n                model.eval()\n            running_loss = 0.0\n            running_corrects = 0\n\n            \n\n            for inputs, labels in tqdm(dataloaders[phase], leave=False):\n                inputs = inputs.to(device)\n                labels = labels.to(device)\n\n                optimizer.zero_grad()\n\n                with torch.set_grad_enabled(phase == 'train'):\n                    outputs = model(inputs)\n                    _, preds = torch.max(outputs, 1)\n                    loss = criterion(outputs, labels)\n\n                    if phase == 'train':\n                        loss.backward()\n                        optimizer.step()\n\n                running_loss += loss.item() * inputs.size(0)\n                running_corrects += torch.sum(preds == labels.data)\n\n            if phase == 'train':\n                scheduler.step()\n\n            epoch_loss = running_loss / dataset_sizes[phase]\n            epoch_acc = running_corrects.double() / dataset_sizes[phase]\n            epoch_data[phase]['loss'].append(epoch_loss)\n            epoch_data[phase]['acc'].append(epoch_acc)\n\n            print(f'{phase} Loss: {epoch_loss:.4f} Acc: {epoch_acc:.4f}')\n\n            if phase == 'val' and epoch_acc > best_acc:\n                best_acc = epoch_acc\n                torch.save(model.state_dict(), best_model_params_path)\n\n        print()\n\n    time_elapsed = time.time() - since\n    print(f'Training complete in {time_elapsed // 60:.0f}m {time_elapsed % 60:.0f}s')\n    print(f'Best val Acc: {best_acc:4f}')\n\n    model.load_state_dict(torch.load(best_model_params_path))\n\n    # Clean up the temporary directory\n    shutil.rmtree(tempdir)\n\n    return model, epoch_data\n","metadata":{"execution":{"iopub.status.busy":"2023-07-25T01:42:58.823345Z","iopub.execute_input":"2023-07-25T01:42:58.823739Z","iopub.status.idle":"2023-07-25T01:42:58.840297Z","shell.execute_reply.started":"2023-07-25T01:42:58.823708Z","shell.execute_reply":"2023-07-25T01:42:58.839256Z"},"trusted":true},"execution_count":22,"outputs":[]},{"cell_type":"code","source":"#train and dataloader code at the end\n#example inputs \nfor img, label in trainloader:\n    example_inputs = img\n    break\nprint(img.shape)\nexample_inputs= example_inputs.to(device)","metadata":{"execution":{"iopub.status.busy":"2023-07-24T09:20:16.585809Z","iopub.execute_input":"2023-07-24T09:20:16.586364Z","iopub.status.idle":"2023-07-24T09:20:16.699154Z","shell.execute_reply.started":"2023-07-24T09:20:16.586332Z","shell.execute_reply":"2023-07-24T09:20:16.698218Z"},"trusted":true},"execution_count":7,"outputs":[{"name":"stdout","text":"torch.Size([32, 3, 224, 224])\n","output_type":"stream"}]},{"cell_type":"code","source":"print(model.features[0])\nprint(model.features[1])","metadata":{"execution":{"iopub.status.busy":"2023-07-24T09:20:16.700406Z","iopub.execute_input":"2023-07-24T09:20:16.701319Z","iopub.status.idle":"2023-07-24T09:20:16.707182Z","shell.execute_reply.started":"2023-07-24T09:20:16.701285Z","shell.execute_reply":"2023-07-24T09:20:16.706287Z"},"trusted":true},"execution_count":8,"outputs":[{"name":"stdout","text":"Conv2dNormActivation(\n  (0): Conv2d(3, 24, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n  (1): BatchNorm2d(24, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n  (2): SiLU(inplace=True)\n)\nSequential(\n  (0): FusedMBConv(\n    (block): Sequential(\n      (0): Conv2dNormActivation(\n        (0): Conv2d(24, 24, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n        (1): BatchNorm2d(24, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n        (2): SiLU(inplace=True)\n      )\n    )\n    (stochastic_depth): StochasticDepth(p=0.0, mode=row)\n  )\n  (1): FusedMBConv(\n    (block): Sequential(\n      (0): Conv2dNormActivation(\n        (0): Conv2d(24, 24, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)\n        (1): BatchNorm2d(24, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n        (2): SiLU(inplace=True)\n      )\n    )\n    (stochastic_depth): StochasticDepth(p=0.005, mode=row)\n  )\n)\n","output_type":"stream"}]},{"cell_type":"code","source":"def get_channel_sizes(model):\n    channel_sizes = []\n\n    for name, module in model.named_modules():\n        if isinstance(module, torch.nn.Conv2d):\n            channel_sizes.append(module.out_channels)\n        \n    print(channel_sizes)\n    print('\\n..................\\nTotal:', len(channel_sizes))\n    return channel_sizes","metadata":{"execution":{"iopub.status.busy":"2023-07-25T01:41:18.842551Z","iopub.execute_input":"2023-07-25T01:41:18.842934Z","iopub.status.idle":"2023-07-25T01:41:18.848781Z","shell.execute_reply.started":"2023-07-25T01:41:18.842905Z","shell.execute_reply":"2023-07-25T01:41:18.847863Z"},"trusted":true},"execution_count":16,"outputs":[]},{"cell_type":"code","source":"#Intial channels\nget_channel_sizes(model)\n","metadata":{"execution":{"iopub.status.busy":"2023-07-24T09:20:16.708481Z","iopub.execute_input":"2023-07-24T09:20:16.709347Z","iopub.status.idle":"2023-07-24T09:20:16.719953Z","shell.execute_reply.started":"2023-07-24T09:20:16.709315Z","shell.execute_reply":"2023-07-24T09:20:16.719023Z"},"trusted":true},"execution_count":9,"outputs":[{"name":"stdout","text":"[24, 24, 24, 96, 48, 192, 48, 192, 48, 192, 48, 192, 64, 256, 64, 256, 64, 256, 64, 256, 256, 16, 256, 128, 512, 512, 32, 512, 128, 512, 512, 32, 512, 128, 512, 512, 32, 512, 128, 512, 512, 32, 512, 128, 512, 512, 32, 512, 128, 768, 768, 32, 768, 160, 960, 960, 40, 960, 160, 960, 960, 40, 960, 160, 960, 960, 40, 960, 160, 960, 960, 40, 960, 160, 960, 960, 40, 960, 160, 960, 960, 40, 960, 160, 960, 960, 40, 960, 160, 960, 960, 40, 960, 160, 960, 960, 40, 960, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1280]\n\n..................\nTotal: 170\n","output_type":"stream"}]},{"cell_type":"markdown","source":"# Defining Pruning","metadata":{}},{"cell_type":"code","source":"# 1. ignore some layers that should not be pruned: the final classifier layer & \n#first 2 MbConv blocks of efficientnet V2 will not be pruned, this is important.\nignored_layers = []\nfor m in model.modules():\n    if isinstance(m, torch.nn.Linear):\n        ignored_layers.append(m) # DO NOT prune the final classifier!\n        \nfor m in model.features[0].modules():  #Ignoring Conv2dNorm\n    if isinstance(m, torchvision.ops.misc.Conv2dNormActivation):\n        ignored_layers.append(m)\n    \nfor m in model.features[0].modules():  #Ignoring 2MB conv blocks\n    if isinstance(m, torchvision.ops.misc.Conv2dNormActivation):\n        ignored_layers.append(m)\n        \nprint(ignored_layers)","metadata":{"execution":{"iopub.status.busy":"2023-07-24T09:20:16.723154Z","iopub.execute_input":"2023-07-24T09:20:16.724064Z","iopub.status.idle":"2023-07-24T09:20:16.733892Z","shell.execute_reply.started":"2023-07-24T09:20:16.724040Z","shell.execute_reply":"2023-07-24T09:20:16.732820Z"},"trusted":true},"execution_count":10,"outputs":[{"name":"stdout","text":"[Linear(in_features=1280, out_features=10, bias=True), Conv2dNormActivation(\n  (0): Conv2d(3, 24, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n  (1): BatchNorm2d(24, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n  (2): SiLU(inplace=True)\n), Conv2dNormActivation(\n  (0): Conv2d(3, 24, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)\n  (1): BatchNorm2d(24, eps=0.001, momentum=0.1, affine=True, track_running_stats=True)\n  (2): SiLU(inplace=True)\n)]\n","output_type":"stream"}]},{"cell_type":"code","source":"imp = tp.importance.LAMPImportance()","metadata":{"execution":{"iopub.status.busy":"2023-07-24T09:20:16.735275Z","iopub.execute_input":"2023-07-24T09:20:16.736120Z","iopub.status.idle":"2023-07-24T09:20:16.745834Z","shell.execute_reply.started":"2023-07-24T09:20:16.736096Z","shell.execute_reply":"2023-07-24T09:20:16.744889Z"},"trusted":true},"execution_count":11,"outputs":[]},{"cell_type":"code","source":"# 2. Pruner initialization\niterative_steps = 10 # You can prune your model to the target sparsity iteratively. #importance parameter\npruner = tp.pruner.MagnitudePruner(\n    model, \n    example_inputs, \n    global_pruning=True, # If False, a uniform sparsity will be assigned to different layers.\n    importance=imp, # importance criterion for parameter selection\n    iterative_steps=iterative_steps, # the number of iterations to achieve target sparsity\n    ch_sparsity=0.8, # Remove 80% channels, \n    ignored_layers=ignored_layers,\n)\n","metadata":{"execution":{"iopub.status.busy":"2023-07-24T09:20:16.747230Z","iopub.execute_input":"2023-07-24T09:20:16.747648Z","iopub.status.idle":"2023-07-24T09:20:21.516655Z","shell.execute_reply.started":"2023-07-24T09:20:16.747617Z","shell.execute_reply":"2023-07-24T09:20:21.515609Z"},"trusted":true},"execution_count":12,"outputs":[]},{"cell_type":"markdown","source":"# Iterative Pruning","metadata":{}},{"cell_type":"code","source":"print('Started channel pruning   ')\nbase_macs, base_nparams = tp.utils.count_ops_and_params(model, example_inputs)\nfor i in range(iterative_steps):\n    pruner.step()\n    print(f'Iteration   {i+1} / {iterative_steps+1}---------------------------')\n    macs, nparams = tp.utils.count_ops_and_params(model, example_inputs)\n    print(model(example_inputs).shape)\n    print(\n        \"  Iter %d/%d, Params: %.2f M => %.2f M\"\n        % (i+1, iterative_steps, base_nparams / 1e6, nparams / 1e6)\n    )\n    print(\n        \"  Iter %d/%d, MACs: %.2f G => %.2f G\"\n        % (i+1, iterative_steps, base_macs / 1e9, macs / 1e9)\n    )\n    print(\"=\"*16)\n    evaluate_model(model)\n    ret_epoch = 3+i\n    print(f'\\n----------- Retraining for {ret_epoch} epochs-------------')\n    \n    \n    #retraining the model \n    criterion = nn.CrossEntropyLoss()\n    # Observe that all parameters are being optimized\n    optimizer_ft = optim.SGD(model.parameters(), lr=0.001, momentum=0.9)\n\n    # Decay LR by a factor of 0.1 every 7 epochs\n    exp_lr_scheduler = lr_scheduler.StepLR(optimizer_ft, step_size=7, gamma=0.1)\n    \n    model, _ = train_model(model, criterion, optimizer_ft, exp_lr_scheduler,\n                       num_epochs= 3 + i )\n    name= (80/10)*(i+1)\n    torch.save(model, f'pruned_model_{name}%.pth')","metadata":{"execution":{"iopub.status.busy":"2023-07-24T09:20:21.518287Z","iopub.execute_input":"2023-07-24T09:20:21.518655Z","iopub.status.idle":"2023-07-24T16:28:15.770062Z","shell.execute_reply.started":"2023-07-24T09:20:21.518621Z","shell.execute_reply":"2023-07-24T16:28:15.769049Z"},"trusted":true},"execution_count":13,"outputs":[{"name":"stdout","text":"Started channel pruning   \nIteration   1 / 11---------------------------\ntorch.Size([32, 10])\n  Iter 1/10, Params: 20.19 M => 18.07 M\n  Iter 1/10, MACs: 2.88 G => 2.80 G\n================\n","output_type":"stream"},{"name":"stderr","text":"100%|██████████| 313/313 [00:33<00:00,  9.25it/s]\n","output_type":"stream"},{"name":"stdout","text":"Evaluation Loss: 0.5805 Acc: 0.8219\n\n----------- Retraining for 3 epochs-------------\nEpoch 1/3\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0847 Acc: 0.9702\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1191 Acc: 0.9658\n\nEpoch 2/3\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0500 Acc: 0.9831\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1127 Acc: 0.9710\n\nEpoch 3/3\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0367 Acc: 0.9879\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1019 Acc: 0.9732\n\nTraining complete in 22m 12s\nBest val Acc: 0.973200\nIteration   2 / 11---------------------------\ntorch.Size([32, 10])\n  Iter 2/10, Params: 20.19 M => 16.01 M\n  Iter 2/10, MACs: 2.88 G => 2.71 G\n================\n","output_type":"stream"},{"name":"stderr","text":"100%|██████████| 313/313 [00:31<00:00, 10.00it/s]\n","output_type":"stream"},{"name":"stdout","text":"Evaluation Loss: 0.7201 Acc: 0.7491\n\n----------- Retraining for 4 epochs-------------\nEpoch 1/4\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0457 Acc: 0.9845\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1058 Acc: 0.9703\n\nEpoch 2/4\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0320 Acc: 0.9893\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1053 Acc: 0.9725\n\nEpoch 3/4\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0245 Acc: 0.9917\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1067 Acc: 0.9740\n\nEpoch 4/4\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0212 Acc: 0.9930\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1074 Acc: 0.9731\n\nTraining complete in 28m 49s\nBest val Acc: 0.974000\nIteration   3 / 11---------------------------\ntorch.Size([32, 10])\n  Iter 3/10, Params: 20.19 M => 14.06 M\n  Iter 3/10, MACs: 2.88 G => 2.61 G\n================\n","output_type":"stream"},{"name":"stderr","text":"100%|██████████| 313/313 [00:30<00:00, 10.29it/s]\n","output_type":"stream"},{"name":"stdout","text":"Evaluation Loss: 0.7596 Acc: 0.7302\n\n----------- Retraining for 5 epochs-------------\nEpoch 1/5\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0386 Acc: 0.9871\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1119 Acc: 0.9698\n\nEpoch 2/5\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0267 Acc: 0.9912\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1108 Acc: 0.9722\n\nEpoch 3/5\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0206 Acc: 0.9932\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1165 Acc: 0.9712\n\nEpoch 4/5\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0159 Acc: 0.9949\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1155 Acc: 0.9718\n\nEpoch 5/5\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0140 Acc: 0.9953\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1135 Acc: 0.9736\n\nTraining complete in 34m 33s\nBest val Acc: 0.973600\nIteration   4 / 11---------------------------\ntorch.Size([32, 10])\n  Iter 4/10, Params: 20.19 M => 12.24 M\n  Iter 4/10, MACs: 2.88 G => 2.50 G\n================\n","output_type":"stream"},{"name":"stderr","text":"100%|██████████| 313/313 [00:29<00:00, 10.58it/s]\n","output_type":"stream"},{"name":"stdout","text":"Evaluation Loss: 0.5882 Acc: 0.8330\n\n----------- Retraining for 6 epochs-------------\nEpoch 1/6\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0366 Acc: 0.9879\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1136 Acc: 0.9695\n\nEpoch 2/6\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0235 Acc: 0.9926\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1134 Acc: 0.9713\n\nEpoch 3/6\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0197 Acc: 0.9935\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1157 Acc: 0.9713\n\nEpoch 4/6\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0160 Acc: 0.9951\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1102 Acc: 0.9731\n\nEpoch 5/6\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0117 Acc: 0.9962\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1155 Acc: 0.9736\n\nEpoch 6/6\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0107 Acc: 0.9964\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1096 Acc: 0.9743\n\nTraining complete in 39m 25s\nBest val Acc: 0.974300\nIteration   5 / 11---------------------------\ntorch.Size([32, 10])\n  Iter 5/10, Params: 20.19 M => 10.51 M\n  Iter 5/10, MACs: 2.88 G => 2.38 G\n================\n","output_type":"stream"},{"name":"stderr","text":"100%|██████████| 313/313 [00:28<00:00, 11.08it/s]\n","output_type":"stream"},{"name":"stdout","text":"Evaluation Loss: 1.1725 Acc: 0.5731\n\n----------- Retraining for 7 epochs-------------\nEpoch 1/7\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0397 Acc: 0.9868\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1255 Acc: 0.9677\n\nEpoch 2/7\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0241 Acc: 0.9925\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1219 Acc: 0.9689\n\nEpoch 3/7\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0178 Acc: 0.9942\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1219 Acc: 0.9713\n\nEpoch 4/7\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0161 Acc: 0.9949\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1154 Acc: 0.9729\n\nEpoch 5/7\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0122 Acc: 0.9961\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1187 Acc: 0.9694\n\nEpoch 6/7\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0096 Acc: 0.9969\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1231 Acc: 0.9709\n\nEpoch 7/7\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0099 Acc: 0.9967\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1335 Acc: 0.9704\n\nTraining complete in 43m 25s\nBest val Acc: 0.972900\nIteration   6 / 11---------------------------\ntorch.Size([32, 10])\n  Iter 6/10, Params: 20.19 M => 8.91 M\n  Iter 6/10, MACs: 2.88 G => 2.25 G\n================\n","output_type":"stream"},{"name":"stderr","text":"100%|██████████| 313/313 [00:27<00:00, 11.41it/s]\n","output_type":"stream"},{"name":"stdout","text":"Evaluation Loss: 1.7625 Acc: 0.4032\n\n----------- Retraining for 8 epochs-------------\nEpoch 1/8\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0571 Acc: 0.9809\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1213 Acc: 0.9663\n\nEpoch 2/8\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0326 Acc: 0.9895\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1225 Acc: 0.9695\n\nEpoch 3/8\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0245 Acc: 0.9918\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1212 Acc: 0.9686\n\nEpoch 4/8\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0206 Acc: 0.9933\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1232 Acc: 0.9685\n\nEpoch 5/8\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0159 Acc: 0.9951\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1288 Acc: 0.9684\n\nEpoch 6/8\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0136 Acc: 0.9956\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1195 Acc: 0.9700\n\nEpoch 7/8\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0127 Acc: 0.9959\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1235 Acc: 0.9687\n\nEpoch 8/8\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0100 Acc: 0.9970\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1179 Acc: 0.9711\n\nTraining complete in 47m 7s\nBest val Acc: 0.971100\nIteration   7 / 11---------------------------\ntorch.Size([32, 10])\n  Iter 7/10, Params: 20.19 M => 7.37 M\n  Iter 7/10, MACs: 2.88 G => 2.13 G\n================\n","output_type":"stream"},{"name":"stderr","text":"100%|██████████| 313/313 [00:26<00:00, 11.81it/s]\n","output_type":"stream"},{"name":"stdout","text":"Evaluation Loss: 4.3306 Acc: 0.1441\n\n----------- Retraining for 9 epochs-------------\nEpoch 1/9\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0660 Acc: 0.9785\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1316 Acc: 0.9603\n\nEpoch 2/9\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0383 Acc: 0.9877\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1260 Acc: 0.9649\n\nEpoch 3/9\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0307 Acc: 0.9901\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1310 Acc: 0.9658\n\nEpoch 4/9\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0248 Acc: 0.9921\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1309 Acc: 0.9649\n\nEpoch 5/9\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0185 Acc: 0.9943\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1274 Acc: 0.9653\n\nEpoch 6/9\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0150 Acc: 0.9959\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1383 Acc: 0.9668\n\nEpoch 7/9\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0150 Acc: 0.9953\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1272 Acc: 0.9682\n\nEpoch 8/9\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0097 Acc: 0.9972\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1225 Acc: 0.9697\n\nEpoch 9/9\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0085 Acc: 0.9974\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1220 Acc: 0.9706\n\nTraining complete in 49m 36s\nBest val Acc: 0.970600\nIteration   8 / 11---------------------------\ntorch.Size([32, 10])\n  Iter 8/10, Params: 20.19 M => 5.88 M\n  Iter 8/10, MACs: 2.88 G => 2.00 G\n================\n","output_type":"stream"},{"name":"stderr","text":"100%|██████████| 313/313 [00:25<00:00, 12.18it/s]\n","output_type":"stream"},{"name":"stdout","text":"Evaluation Loss: 2.9468 Acc: 0.1785\n\n----------- Retraining for 10 epochs-------------\nEpoch 1/10\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0940 Acc: 0.9691\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1429 Acc: 0.9563\n\nEpoch 2/10\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0548 Acc: 0.9819\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1294 Acc: 0.9622\n\nEpoch 3/10\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0391 Acc: 0.9872\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1479 Acc: 0.9584\n\nEpoch 4/10\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0316 Acc: 0.9899\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1365 Acc: 0.9625\n\nEpoch 5/10\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0273 Acc: 0.9912\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1415 Acc: 0.9632\n\nEpoch 6/10\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0195 Acc: 0.9937\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1380 Acc: 0.9633\n\nEpoch 7/10\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0202 Acc: 0.9939\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1421 Acc: 0.9619\n\nEpoch 8/10\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0140 Acc: 0.9959\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1308 Acc: 0.9657\n\nEpoch 9/10\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0117 Acc: 0.9966\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1303 Acc: 0.9661\n\nEpoch 10/10\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0101 Acc: 0.9972\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1305 Acc: 0.9666\n\nTraining complete in 51m 22s\nBest val Acc: 0.966600\nIteration   9 / 11---------------------------\ntorch.Size([32, 10])\n  Iter 9/10, Params: 20.19 M => 4.41 M\n  Iter 9/10, MACs: 2.88 G => 1.87 G\n================\n","output_type":"stream"},{"name":"stderr","text":"100%|██████████| 313/313 [00:24<00:00, 12.97it/s]\n","output_type":"stream"},{"name":"stdout","text":"Evaluation Loss: 4.1527 Acc: 0.1002\n\n----------- Retraining for 11 epochs-------------\nEpoch 1/11\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.1611 Acc: 0.9461\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1793 Acc: 0.9417\n\nEpoch 2/11\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0877 Acc: 0.9710\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1484 Acc: 0.9536\n\nEpoch 3/11\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0615 Acc: 0.9798\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1643 Acc: 0.9512\n\nEpoch 4/11\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0495 Acc: 0.9844\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1587 Acc: 0.9539\n\nEpoch 5/11\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0425 Acc: 0.9864\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1618 Acc: 0.9549\n\nEpoch 6/11\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0366 Acc: 0.9878\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1603 Acc: 0.9555\n\nEpoch 7/11\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0305 Acc: 0.9903\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1555 Acc: 0.9581\n\nEpoch 8/11\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0207 Acc: 0.9942\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1473 Acc: 0.9599\n\nEpoch 9/11\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0169 Acc: 0.9952\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1410 Acc: 0.9619\n\nEpoch 10/11\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0157 Acc: 0.9959\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1415 Acc: 0.9629\n\nEpoch 11/11\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0138 Acc: 0.9961\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1459 Acc: 0.9613\n\nTraining complete in 52m 45s\nBest val Acc: 0.962900\nIteration   10 / 11---------------------------\ntorch.Size([32, 10])\n  Iter 10/10, Params: 20.19 M => 2.93 M\n  Iter 10/10, MACs: 2.88 G => 1.65 G\n================\n","output_type":"stream"},{"name":"stderr","text":"100%|██████████| 313/313 [00:22<00:00, 13.94it/s]\n","output_type":"stream"},{"name":"stdout","text":"Evaluation Loss: 10.9966 Acc: 0.1000\n\n----------- Retraining for 12 epochs-------------\nEpoch 1/12\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.2560 Acc: 0.9178\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1956 Acc: 0.9350\n\nEpoch 2/12\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.1412 Acc: 0.9540\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1709 Acc: 0.9436\n\nEpoch 3/12\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.1083 Acc: 0.9645\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1718 Acc: 0.9455\n\nEpoch 4/12\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0850 Acc: 0.9722\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1721 Acc: 0.9464\n\nEpoch 5/12\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0720 Acc: 0.9767\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1735 Acc: 0.9478\n\nEpoch 6/12\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0620 Acc: 0.9802\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1805 Acc: 0.9482\n\nEpoch 7/12\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0523 Acc: 0.9829\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1718 Acc: 0.9516\n\nEpoch 8/12\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0361 Acc: 0.9888\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1544 Acc: 0.9553\n\nEpoch 9/12\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0287 Acc: 0.9917\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1570 Acc: 0.9551\n\nEpoch 10/12\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0272 Acc: 0.9923\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1592 Acc: 0.9537\n\nEpoch 11/12\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0230 Acc: 0.9936\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1542 Acc: 0.9565\n\nEpoch 12/12\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0228 Acc: 0.9936\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1548 Acc: 0.9573\n\nTraining complete in 53m 48s\nBest val Acc: 0.957300\n","output_type":"stream"}]},{"cell_type":"code","source":"","metadata":{"execution":{"iopub.status.busy":"2023-07-24T16:28:15.772846Z","iopub.execute_input":"2023-07-24T16:28:15.773388Z","iopub.status.idle":"2023-07-24T16:28:15.778577Z","shell.execute_reply.started":"2023-07-24T16:28:15.773352Z","shell.execute_reply":"2023-07-24T16:28:15.777594Z"},"trusted":true},"execution_count":14,"outputs":[]},{"cell_type":"code","source":"import numpy as np\nchannels_init = [24, 24, 24, 96, 48, 192, 48, 192, 48, 192, 48, 192, 64, 256, 64, 256, 64, 256, 64, 256, 256, 16, 256, 128, 512, 512, 32, 512, 128, 512, 512, 32, 512, 128, 512, 512, 32, 512, 128, 512, 512, 32, 512, 128, 512, 512, 32, 512, 128, 768, 768, 32, 768, 160, 960, 960, 40, 960, 160, 960, 960, 40, 960, 160, 960, 960, 40, 960, 160, 960, 960, 40, 960, 160, 960, 960, 40, 960, 160, 960, 960, 40, 960, 160, 960, 960, 40, 960, 160, 960, 960, 40, 960, 160, 960, 960, 40, 960, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1280]\nchannels_prune = get_channel_sizes(model)\n\nchannels_init = np.array(channels_init)\nchannels_prune = np.array(channels_prune)\ntotal_ch_initialy = np.sum(channels_init)\ntotal_ch_prune = np.sum(channels_prune)\n\nper =  (total_ch_initialy-total_ch_prune)/total_ch_initialy\n\nprint(f'Pruned percentage {per} %')\n\n\n","metadata":{"execution":{"iopub.status.busy":"2023-07-24T16:29:02.470880Z","iopub.execute_input":"2023-07-24T16:29:02.471248Z","iopub.status.idle":"2023-07-24T16:29:02.489048Z","shell.execute_reply.started":"2023-07-24T16:29:02.471218Z","shell.execute_reply":"2023-07-24T16:29:02.488125Z"},"trusted":true},"execution_count":16,"outputs":[{"name":"stdout","text":"[24, 24, 24, 95, 48, 153, 48, 191, 48, 188, 48, 192, 64, 227, 64, 242, 64, 226, 64, 203, 203, 16, 203, 128, 145, 145, 32, 145, 128, 127, 127, 32, 127, 128, 150, 150, 14, 150, 128, 142, 142, 9, 142, 128, 144, 144, 2, 144, 128, 168, 168, 32, 168, 160, 118, 118, 40, 118, 160, 128, 128, 28, 128, 160, 133, 133, 15, 133, 160, 135, 135, 7, 135, 160, 160, 160, 2, 160, 160, 114, 114, 12, 114, 160, 153, 153, 5, 153, 160, 132, 132, 3, 132, 160, 238, 238, 40, 238, 256, 119, 119, 20, 119, 256, 143, 143, 49, 143, 256, 136, 136, 60, 136, 256, 131, 131, 44, 131, 256, 123, 123, 44, 123, 256, 120, 120, 64, 120, 256, 117, 117, 64, 117, 256, 139, 139, 48, 139, 256, 169, 169, 64, 169, 256, 132, 132, 64, 132, 256, 112, 112, 64, 112, 256, 120, 120, 61, 120, 256, 132, 132, 64, 132, 256, 139, 139, 64, 139, 256, 296]\n\n..................\nTotal: 170\nPruned percentage 0.8028413955479452 %\n","output_type":"stream"}]},{"cell_type":"code","source":"import matplotlib.pyplot as plt\nx = range(1, 171)\n\n# Create a figure and axis\n\n# fig, ax = plt.subplots()\nplt.figure(figsize=(15,5))\n\n# Plot the first array with a blue curve, labeled with a marker\nplt.plot(x, channels_init, color='blue')\n\n# Plot the second array with a red curve, labeled with a marker\nplt.plot(x, channels_prune, color='red')\n\n# Add legend\nplt.legend()\n\n# Set labels for x and y axes\nplt.ylabel('No. of Channels')\nplt.xlabel('Layers')\n\n# Show the plot\nplt.show()\n\n# Example usage:\nimport random\n","metadata":{"execution":{"iopub.status.busy":"2023-07-24T16:35:59.561014Z","iopub.execute_input":"2023-07-24T16:35:59.561771Z","iopub.status.idle":"2023-07-24T16:35:59.856483Z","shell.execute_reply.started":"2023-07-24T16:35:59.561730Z","shell.execute_reply":"2023-07-24T16:35:59.855640Z"},"trusted":true},"execution_count":31,"outputs":[{"output_type":"display_data","data":{"text/plain":"<Figure size 1500x500 with 1 Axes>","image/png":"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"},"metadata":{}}]},{"cell_type":"markdown","source":"# Model Sizes","metadata":{}},{"cell_type":"code","source":"import os\n\nfile_size_bytes = os.path.getsize('/kaggle/input/pruning-effficientnet-v2-s-on-cifar10/best_model_9791.pt')\nfile_size_mb = file_size_bytes / (1024 * 1024)\nprint(f\"Model file size before pruning: {file_size_mb:.2f} MB\")\n\nfile_size_bytes = os.path.getsize('/kaggle/working/pruned_model_80.0%.pth')\nfile_size_mb = file_size_bytes / (1024 * 1024)\nprint(f\"Model file size after pruning: {file_size_mb:.2f} MB\")","metadata":{"execution":{"iopub.status.busy":"2023-07-24T16:50:21.875872Z","iopub.execute_input":"2023-07-24T16:50:21.876246Z","iopub.status.idle":"2023-07-24T16:50:21.884199Z","shell.execute_reply.started":"2023-07-24T16:50:21.876216Z","shell.execute_reply":"2023-07-24T16:50:21.883272Z"},"trusted":true},"execution_count":46,"outputs":[{"name":"stdout","text":"Model file size before pruning: 77.99 MB\nModel file size after pruning: 11.68 MB\n","output_type":"stream"}]},{"cell_type":"markdown","source":"# Few More tests \nUpto this point 80% channels are pruned, Let's try pruning 20% channels from remaining channels","metadata":{}},{"cell_type":"code","source":"iterative_steps = 1 # You can prune your model to the target sparsity iteratively. #importance parameter\npruner = tp.pruner.MagnitudePruner(\n    model, \n    example_inputs, \n    global_pruning=True, # If False, a uniform sparsity will be assigned to different layers.\n    importance=imp, # importance criterion for parameter selection\n    iterative_steps=iterative_steps, # the number of iterations to achieve target sparsity\n    ch_sparsity=0.2, # Remove 80% channels, \n    ignored_layers=ignored_layers,\n)\n","metadata":{"execution":{"iopub.status.busy":"2023-07-24T17:07:55.218201Z","iopub.execute_input":"2023-07-24T17:07:55.218894Z","iopub.status.idle":"2023-07-24T17:07:55.368716Z","shell.execute_reply.started":"2023-07-24T17:07:55.218860Z","shell.execute_reply":"2023-07-24T17:07:55.367802Z"},"trusted":true},"execution_count":54,"outputs":[]},{"cell_type":"markdown","source":"","metadata":{}},{"cell_type":"code","source":"for i in range(iterative_steps):\n    pruner.step()\n    print(f'Iteration   {i+1} / {iterative_steps}---------------------------')\n    macs, nparams = tp.utils.count_ops_and_params(model, example_inputs)\n    print(model(example_inputs).shape)\n    print(\n        \"  Iter %d/%d, Params: %.2f M => %.2f M\"\n        % (i+1, iterative_steps, base_nparams / 1e6, nparams / 1e6)\n    )\n    print(\n        \"  Iter %d/%d, MACs: %.2f G => %.2f G\"\n        % (i+1, iterative_steps, base_macs / 1e9, macs / 1e9)\n    )\n    print(\"=\"*16)","metadata":{"execution":{"iopub.status.busy":"2023-07-24T16:52:54.760340Z","iopub.execute_input":"2023-07-24T16:52:54.760728Z","iopub.status.idle":"2023-07-24T16:52:55.729113Z","shell.execute_reply.started":"2023-07-24T16:52:54.760699Z","shell.execute_reply":"2023-07-24T16:52:55.728147Z"},"trusted":true},"execution_count":48,"outputs":[{"name":"stdout","text":"Iteration   1 / 2---------------------------\ntorch.Size([32, 10])\n  Iter 1/1, Params: 20.19 M => 2.01 M\n  Iter 1/1, MACs: 2.88 G => 1.35 G\n================\n","output_type":"stream"}]},{"cell_type":"code","source":"evaluate_model(model)","metadata":{"execution":{"iopub.status.busy":"2023-07-24T16:54:36.195749Z","iopub.execute_input":"2023-07-24T16:54:36.196130Z","iopub.status.idle":"2023-07-24T16:54:57.530659Z","shell.execute_reply.started":"2023-07-24T16:54:36.196101Z","shell.execute_reply":"2023-07-24T16:54:57.529751Z"},"trusted":true},"execution_count":49,"outputs":[{"name":"stderr","text":"100%|██████████| 313/313 [00:21<00:00, 14.69it/s]\n","output_type":"stream"},{"name":"stdout","text":"Evaluation Loss: 11.6117 Acc: 0.1288\n","output_type":"stream"},{"execution_count":49,"output_type":"execute_result","data":{"text/plain":"(11.61174298248291, tensor(0.1288, device='cuda:0', dtype=torch.float64))"},"metadata":{}}]},{"cell_type":"code","source":"    criterion = nn.CrossEntropyLoss()\n    # Observe that all parameters are being optimized\n    optimizer_ft = optim.SGD(model.parameters(), lr=0.001, momentum=0.9)\n\n    # Decay LR by a factor of 0.1 every 7 epochs\n    exp_lr_scheduler = lr_scheduler.StepLR(optimizer_ft, step_size=7, gamma=0.1)\n    \n    model, _ = train_model(model, criterion, optimizer_ft, exp_lr_scheduler,\n                       num_epochs= 2 )\n    name= 84 #this is total pruned per\n#  torch.save(model, f'pruned_model_{name}%.pth')","metadata":{"execution":{"iopub.status.busy":"2023-07-24T16:56:22.064381Z","iopub.execute_input":"2023-07-24T16:56:22.065065Z","iopub.status.idle":"2023-07-24T17:04:59.939601Z","shell.execute_reply.started":"2023-07-24T16:56:22.065031Z","shell.execute_reply":"2023-07-24T17:04:59.938640Z"},"trusted":true},"execution_count":50,"outputs":[{"name":"stdout","text":"Epoch 1/2\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.2796 Acc: 0.9081\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.2133 Acc: 0.9289\n\nEpoch 2/2\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.1680 Acc: 0.9453\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.2088 Acc: 0.9326\n\nTraining complete in 8m 38s\nBest val Acc: 0.932600\n","output_type":"stream"}]},{"cell_type":"code","source":"","metadata":{"execution":{"iopub.status.busy":"2023-07-24T17:06:07.653303Z","iopub.execute_input":"2023-07-24T17:06:07.653721Z","iopub.status.idle":"2023-07-24T17:06:07.670957Z","shell.execute_reply.started":"2023-07-24T17:06:07.653690Z","shell.execute_reply":"2023-07-24T17:06:07.669868Z"},"trusted":true},"execution_count":51,"outputs":[{"name":"stdout","text":"[24, 24, 24, 91, 48, 111, 48, 161, 48, 167, 48, 180, 64, 172, 64, 187, 64, 175, 64, 150, 150, 16, 150, 128, 101, 101, 32, 101, 128, 93, 93, 32, 93, 128, 108, 108, 14, 108, 128, 99, 99, 9, 99, 128, 115, 115, 2, 115, 128, 132, 132, 32, 132, 147, 90, 90, 40, 90, 147, 92, 92, 27, 92, 147, 87, 87, 15, 87, 147, 100, 100, 6, 100, 147, 117, 117, 1, 117, 147, 102, 102, 12, 102, 147, 121, 121, 5, 121, 147, 99, 99, 2, 99, 147, 178, 178, 40, 178, 195, 86, 86, 17, 86, 195, 106, 106, 46, 106, 195, 118, 118, 54, 118, 195, 87, 87, 42, 87, 195, 79, 79, 42, 79, 195, 99, 99, 62, 99, 195, 89, 89, 62, 89, 195, 101, 101, 47, 101, 195, 129, 129, 64, 129, 195, 87, 87, 62, 87, 195, 82, 82, 64, 82, 195, 84, 84, 60, 84, 195, 90, 90, 64, 90, 195, 107, 107, 64, 107, 195, 211]\n\n..................\nTotal: 170\nPruned percentage 0.8447310216894978 %\n","output_type":"stream"}]},{"cell_type":"code","source":"torch.save(model, f'pruned_model_{name}%.pth')","metadata":{"execution":{"iopub.status.busy":"2023-07-24T17:07:29.583750Z","iopub.execute_input":"2023-07-24T17:07:29.584132Z","iopub.status.idle":"2023-07-24T17:07:29.726740Z","shell.execute_reply.started":"2023-07-24T17:07:29.584104Z","shell.execute_reply":"2023-07-24T17:07:29.725509Z"},"trusted":true},"execution_count":52,"outputs":[]},{"cell_type":"markdown","source":"Seems like we try pruning 20% more and see the results","metadata":{}},{"cell_type":"code","source":"for i in range(iterative_steps):\n    pruner.step()\n    print(f'Iteration   {i+1} / {iterative_steps}---------------------------')\n    macs, nparams = tp.utils.count_ops_and_params(model, example_inputs)\n    print(model(example_inputs).shape)\n    print(\n        \"  Iter %d/%d, Params: %.2f M => %.2f M\"\n        % (i+1, iterative_steps, base_nparams / 1e6, nparams / 1e6)\n    )\n    print(\n        \"  Iter %d/%d, MACs: %.2f G => %.2f G\"\n        % (i+1, iterative_steps, base_macs / 1e9, macs / 1e9)\n    )\n    print(\"=\"*16)","metadata":{"execution":{"iopub.status.busy":"2023-07-24T17:08:03.172285Z","iopub.execute_input":"2023-07-24T17:08:03.172654Z","iopub.status.idle":"2023-07-24T17:08:03.872821Z","shell.execute_reply.started":"2023-07-24T17:08:03.172624Z","shell.execute_reply":"2023-07-24T17:08:03.871806Z"},"trusted":true},"execution_count":55,"outputs":[{"name":"stdout","text":"Iteration   1 / 1---------------------------\ntorch.Size([32, 10])\n  Iter 1/1, Params: 20.19 M => 1.36 M\n  Iter 1/1, MACs: 2.88 G => 1.07 G\n================\n","output_type":"stream"}]},{"cell_type":"code","source":"evaluate_model(model)","metadata":{"execution":{"iopub.status.busy":"2023-07-24T17:08:18.964536Z","iopub.execute_input":"2023-07-24T17:08:18.965265Z","iopub.status.idle":"2023-07-24T17:08:40.855444Z","shell.execute_reply.started":"2023-07-24T17:08:18.965226Z","shell.execute_reply":"2023-07-24T17:08:40.854509Z"},"trusted":true},"execution_count":56,"outputs":[{"name":"stderr","text":"100%|██████████| 313/313 [00:21<00:00, 14.31it/s]","output_type":"stream"},{"name":"stdout","text":"Evaluation Loss: 11.3746 Acc: 0.1000\n","output_type":"stream"},{"name":"stderr","text":"\n","output_type":"stream"},{"execution_count":56,"output_type":"execute_result","data":{"text/plain":"(11.374557704162598, tensor(0.1000, device='cuda:0', dtype=torch.float64))"},"metadata":{}}]},{"cell_type":"code","source":"criterion = nn.CrossEntropyLoss()\n# Observe that all parameters are being optimized\noptimizer_ft = optim.SGD(model.parameters(), lr=0.001, momentum=0.9)\n\n# Decay LR by a factor of 0.1 every 7 epochs\nexp_lr_scheduler = lr_scheduler.StepLR(optimizer_ft, step_size=7, gamma=0.1)\n\nmodel, _ = train_model(model, criterion, optimizer_ft, exp_lr_scheduler,\n                   num_epochs= 20 )\nname= 87 #this is total pruned per","metadata":{"execution":{"iopub.status.busy":"2023-07-24T17:09:17.938298Z","iopub.execute_input":"2023-07-24T17:09:17.938975Z","iopub.status.idle":"2023-07-24T18:31:34.736178Z","shell.execute_reply.started":"2023-07-24T17:09:17.938933Z","shell.execute_reply":"2023-07-24T18:31:34.735148Z"},"trusted":true},"execution_count":57,"outputs":[{"name":"stdout","text":"Epoch 1/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.4493 Acc: 0.8517\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.2748 Acc: 0.9079\n\nEpoch 2/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.2780 Acc: 0.9084\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.2347 Acc: 0.9190\n\nEpoch 3/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.2258 Acc: 0.9251\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.2449 Acc: 0.9183\n\nEpoch 4/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.1914 Acc: 0.9369\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.2089 Acc: 0.9290\n\nEpoch 5/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.1618 Acc: 0.9461\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.2192 Acc: 0.9280\n\nEpoch 6/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.1444 Acc: 0.9527\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.2076 Acc: 0.9326\n\nEpoch 7/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.1243 Acc: 0.9600\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.2097 Acc: 0.9342\n\nEpoch 8/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0869 Acc: 0.9728\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1874 Acc: 0.9428\n\nEpoch 9/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0750 Acc: 0.9770\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1856 Acc: 0.9431\n\nEpoch 10/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0666 Acc: 0.9792\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1857 Acc: 0.9426\n\nEpoch 11/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0635 Acc: 0.9806\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1886 Acc: 0.9429\n\nEpoch 12/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0618 Acc: 0.9810\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1867 Acc: 0.9437\n\nEpoch 13/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0573 Acc: 0.9826\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1841 Acc: 0.9466\n\nEpoch 14/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0553 Acc: 0.9826\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1852 Acc: 0.9444\n\nEpoch 15/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0525 Acc: 0.9842\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1828 Acc: 0.9458\n\nEpoch 16/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0492 Acc: 0.9854\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1852 Acc: 0.9469\n\nEpoch 17/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0509 Acc: 0.9849\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1843 Acc: 0.9479\n\nEpoch 18/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0507 Acc: 0.9842\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1840 Acc: 0.9461\n\nEpoch 19/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0499 Acc: 0.9849\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1865 Acc: 0.9456\n\nEpoch 20/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0484 Acc: 0.9853\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1839 Acc: 0.9466\n\nTraining complete in 82m 17s\nBest val Acc: 0.947900\n","output_type":"stream"}]},{"cell_type":"code","source":"torch.save(model, f'pruned_model_{name}%.pth')","metadata":{"execution":{"iopub.status.busy":"2023-07-24T18:31:34.877106Z","iopub.execute_input":"2023-07-24T18:31:34.877681Z","iopub.status.idle":"2023-07-24T18:31:35.015375Z","shell.execute_reply.started":"2023-07-24T18:31:34.877655Z","shell.execute_reply":"2023-07-24T18:31:35.014488Z"},"trusted":true},"execution_count":59,"outputs":[]},{"cell_type":"markdown","source":"# Final Pruned model visualization","metadata":{}},{"cell_type":"code","source":"import numpy as np\nchannels_init = [24, 24, 24, 96, 48, 192, 48, 192, 48, 192, 48, 192, 64, 256, 64, 256, 64, 256, 64, 256, 256, 16, 256, 128, 512, 512, 32, 512, 128, 512, 512, 32, 512, 128, 512, 512, 32, 512, 128, 512, 512, 32, 512, 128, 512, 512, 32, 512, 128, 768, 768, 32, 768, 160, 960, 960, 40, 960, 160, 960, 960, 40, 960, 160, 960, 960, 40, 960, 160, 960, 960, 40, 960, 160, 960, 960, 40, 960, 160, 960, 960, 40, 960, 160, 960, 960, 40, 960, 160, 960, 960, 40, 960, 160, 960, 960, 40, 960, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1536, 1536, 64, 1536, 256, 1280]\nchannels_prune = get_channel_sizes(model)\n\nchannels_init = np.array(channels_init)\nchannels_prune = np.array(channels_prune)\ntotal_ch_initialy = np.sum(channels_init)\ntotal_ch_prune = np.sum(channels_prune)\n\nper =  (total_ch_initialy-total_ch_prune)/total_ch_initialy\n\nprint(f'Pruned percentage {per} %')\n","metadata":{"execution":{"iopub.status.busy":"2023-07-25T01:41:28.367404Z","iopub.execute_input":"2023-07-25T01:41:28.367788Z","iopub.status.idle":"2023-07-25T01:41:28.386171Z","shell.execute_reply.started":"2023-07-25T01:41:28.367757Z","shell.execute_reply":"2023-07-25T01:41:28.384905Z"},"trusted":true},"execution_count":17,"outputs":[{"name":"stdout","text":"[24, 24, 24, 77, 48, 84, 48, 127, 48, 134, 48, 145, 64, 132, 64, 139, 64, 134, 64, 109, 109, 16, 109, 128, 82, 82, 32, 82, 128, 78, 78, 32, 78, 128, 88, 88, 14, 88, 128, 78, 78, 9, 78, 128, 88, 88, 2, 88, 128, 98, 98, 32, 98, 113, 64, 64, 40, 64, 113, 66, 66, 26, 66, 113, 60, 60, 15, 60, 113, 68, 68, 6, 68, 113, 88, 88, 1, 88, 113, 78, 78, 12, 78, 113, 94, 94, 5, 94, 113, 65, 65, 2, 65, 113, 128, 128, 40, 128, 148, 71, 71, 14, 71, 148, 77, 77, 42, 77, 148, 100, 100, 44, 100, 148, 70, 70, 40, 70, 148, 61, 61, 35, 61, 148, 75, 75, 54, 75, 148, 57, 57, 58, 57, 148, 70, 70, 45, 70, 148, 91, 91, 64, 91, 148, 76, 76, 58, 76, 148, 59, 59, 61, 59, 148, 61, 61, 56, 61, 148, 64, 64, 63, 64, 148, 84, 84, 63, 84, 148, 159]\n\n..................\nTotal: 170\nPruned percentage 0.8779876569634704 %\n","output_type":"stream"}]},{"cell_type":"code","source":"import matplotlib.pyplot as plt\nx = range(1, 171)\n\n# Create a figure and axis\n\n# fig, ax = plt.subplots()\nplt.figure(figsize=(15,5))\n\n# Plot the first array with a blue curve, labeled with a marker\nplt.plot(x, channels_init, color='blue')\n\n# Plot the second array with a red curve, labeled with a marker\nplt.plot(x, channels_prune, color='red')\n\n# Add legend\nplt.legend()\n\n# Set labels for x and y axes\nplt.ylabel('No. of Channels')\nplt.xlabel('Layers')\n\n# Show the plot\nplt.show()\n\n# Example usage:\nimport random\n","metadata":{"execution":{"iopub.status.busy":"2023-07-25T01:41:31.462408Z","iopub.execute_input":"2023-07-25T01:41:31.462769Z","iopub.status.idle":"2023-07-25T01:41:31.906017Z","shell.execute_reply.started":"2023-07-25T01:41:31.462741Z","shell.execute_reply":"2023-07-25T01:41:31.905017Z"},"trusted":true},"execution_count":18,"outputs":[{"output_type":"display_data","data":{"text/plain":"<Figure size 1500x500 with 1 Axes>","image/png":"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"},"metadata":{}}]},{"cell_type":"code","source":"","metadata":{"execution":{"iopub.status.busy":"2023-07-24T18:31:35.034782Z","iopub.execute_input":"2023-07-24T18:31:35.035114Z","iopub.status.idle":"2023-07-24T18:31:35.274334Z","shell.execute_reply.started":"2023-07-24T18:31:35.035081Z","shell.execute_reply":"2023-07-24T18:31:35.273277Z"},"trusted":true},"execution_count":61,"outputs":[{"name":"stdout","text":"Iteration   1 / 1---------------------------\ntorch.Size([32, 10])\n  Iter 1/1, Params: 20.19 M => 1.36 M\n  Iter 1/1, MACs: 2.88 G => 1.07 G\n================\n","output_type":"stream"}]},{"cell_type":"markdown","source":"# Measuring Inference time and model size\n","metadata":{}},{"cell_type":"code","source":"! pip install nn-meter\n!nn-meter --list-predictors","metadata":{"execution":{"iopub.status.busy":"2023-07-25T01:33:32.783531Z","iopub.execute_input":"2023-07-25T01:33:32.784340Z","iopub.status.idle":"2023-07-25T01:33:52.581318Z","shell.execute_reply.started":"2023-07-25T01:33:32.784306Z","shell.execute_reply":"2023-07-25T01:33:52.579927Z"},"trusted":true},"execution_count":1,"outputs":[{"name":"stdout","text":"Collecting nn-meter\n  Downloading nn_meter-2.0-py3-none-any.whl (132 kB)\n\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m132.5/132.5 kB\u001b[0m \u001b[31m3.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0ma \u001b[36m0:00:01\u001b[0m\n\u001b[?25hRequirement already satisfied: numpy in /opt/conda/lib/python3.10/site-packages (from nn-meter) (1.23.5)\nRequirement already satisfied: pandas in /opt/conda/lib/python3.10/site-packages (from nn-meter) (1.5.3)\nRequirement already satisfied: tqdm in /opt/conda/lib/python3.10/site-packages (from nn-meter) (4.64.1)\nRequirement already satisfied: networkx in /opt/conda/lib/python3.10/site-packages (from nn-meter) (3.1)\nRequirement already satisfied: requests in /opt/conda/lib/python3.10/site-packages (from nn-meter) (2.28.2)\nRequirement already satisfied: protobuf in /opt/conda/lib/python3.10/site-packages (from nn-meter) (3.20.3)\nRequirement already satisfied: PyYAML in /opt/conda/lib/python3.10/site-packages (from nn-meter) (5.4.1)\nRequirement already satisfied: scikit-learn in /opt/conda/lib/python3.10/site-packages (from nn-meter) (1.2.2)\nRequirement already satisfied: packaging in /opt/conda/lib/python3.10/site-packages (from nn-meter) (21.3)\nCollecting jsonlines (from nn-meter)\n  Downloading jsonlines-3.1.0-py3-none-any.whl (8.6 kB)\nRequirement already satisfied: attrs>=19.2.0 in /opt/conda/lib/python3.10/site-packages (from jsonlines->nn-meter) (23.1.0)\nRequirement already satisfied: pyparsing!=3.0.5,>=2.0.2 in /opt/conda/lib/python3.10/site-packages (from packaging->nn-meter) (3.0.9)\nRequirement already satisfied: python-dateutil>=2.8.1 in /opt/conda/lib/python3.10/site-packages (from pandas->nn-meter) (2.8.2)\nRequirement already satisfied: pytz>=2020.1 in /opt/conda/lib/python3.10/site-packages (from pandas->nn-meter) (2023.3)\nRequirement already satisfied: charset-normalizer<4,>=2 in /opt/conda/lib/python3.10/site-packages (from requests->nn-meter) (2.1.1)\nRequirement already satisfied: idna<4,>=2.5 in /opt/conda/lib/python3.10/site-packages (from requests->nn-meter) (3.4)\nRequirement already satisfied: urllib3<1.27,>=1.21.1 in /opt/conda/lib/python3.10/site-packages (from requests->nn-meter) (1.26.15)\nRequirement already satisfied: certifi>=2017.4.17 in /opt/conda/lib/python3.10/site-packages (from requests->nn-meter) (2023.5.7)\nRequirement already satisfied: scipy>=1.3.2 in /opt/conda/lib/python3.10/site-packages (from scikit-learn->nn-meter) (1.10.1)\nRequirement already satisfied: joblib>=1.1.1 in /opt/conda/lib/python3.10/site-packages (from scikit-learn->nn-meter) (1.2.0)\nRequirement already satisfied: threadpoolctl>=2.0.0 in /opt/conda/lib/python3.10/site-packages (from scikit-learn->nn-meter) (3.1.0)\nRequirement already satisfied: six>=1.5 in /opt/conda/lib/python3.10/site-packages (from python-dateutil>=2.8.1->pandas->nn-meter) (1.16.0)\nInstalling collected packages: jsonlines, nn-meter\nSuccessfully installed jsonlines-3.1.0 nn-meter-2.0\n\u001b[33mWARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv\u001b[0m\u001b[33m\n\u001b[0mTraceback (most recent call last):\n  File \"/opt/conda/lib/python3.10/site-packages/pkg_resources/__init__.py\", line 3031, in _dep_map\n    return self.__dep_map\n  File \"/opt/conda/lib/python3.10/site-packages/pkg_resources/__init__.py\", line 2828, in __getattr__\n    raise AttributeError(attr)\nAttributeError: _DistInfoDistribution__dep_map\n\nDuring handling of the above exception, another exception occurred:\n\nTraceback (most recent call last):\n  File \"/opt/conda/lib/python3.10/site-packages/pkg_resources/__init__.py\", line 3022, in _parsed_pkg_info\n    return self._pkg_info\n  File \"/opt/conda/lib/python3.10/site-packages/pkg_resources/__init__.py\", line 2828, in __getattr__\n    raise AttributeError(attr)\nAttributeError: _pkg_info. Did you mean: 'egg_info'?\n\nDuring handling of the above exception, another exception occurred:\n\nTraceback (most recent call last):\n  File \"/opt/conda/bin/nn-meter\", line 5, in <module>\n    from nn_meter.utils.nn_meter_cli.interface import nn_meter_cli\n  File \"/opt/conda/lib/python3.10/site-packages/nn_meter/__init__.py\", line 6, in <module>\n    __version__ = pkg_resources.require(\"nn-meter\")[0].version\n  File \"/opt/conda/lib/python3.10/site-packages/pkg_resources/__init__.py\", line 891, in require\n    needed = self.resolve(parse_requirements(requirements))\n  File \"/opt/conda/lib/python3.10/site-packages/pkg_resources/__init__.py\", line 785, in resolve\n    new_requirements = dist.requires(req.extras)[::-1]\n  File \"/opt/conda/lib/python3.10/site-packages/pkg_resources/__init__.py\", line 2749, in requires\n    dm = self._dep_map\n  File \"/opt/conda/lib/python3.10/site-packages/pkg_resources/__init__.py\", line 3033, in _dep_map\n    self.__dep_map = self._compute_dependencies()\n  File \"/opt/conda/lib/python3.10/site-packages/pkg_resources/__init__.py\", line 3042, in _compute_dependencies\n    for req in self._parsed_pkg_info.get_all('Requires-Dist') or []:\n  File \"/opt/conda/lib/python3.10/site-packages/pkg_resources/__init__.py\", line 3024, in _parsed_pkg_info\n    metadata = self.get_metadata(self.PKG_INFO)\n  File \"/opt/conda/lib/python3.10/site-packages/pkg_resources/__init__.py\", line 1412, in get_metadata\n    value = self._get(path)\n  File \"/opt/conda/lib/python3.10/site-packages/pkg_resources/__init__.py\", line 1616, in _get\n    with open(path, 'rb') as stream:\nFileNotFoundError: [Errno 2] No such file or directory: '/opt/conda/lib/python3.10/site-packages/numpy-1.24.3.dist-info/METADATA'\n","output_type":"stream"}]},{"cell_type":"code","source":"!pip install onnx\n!pip install onnx-simplifier","metadata":{"execution":{"iopub.status.busy":"2023-07-25T01:37:29.490068Z","iopub.execute_input":"2023-07-25T01:37:29.490500Z","iopub.status.idle":"2023-07-25T01:37:54.354234Z","shell.execute_reply.started":"2023-07-25T01:37:29.490472Z","shell.execute_reply":"2023-07-25T01:37:54.352894Z"},"trusted":true},"execution_count":9,"outputs":[{"name":"stdout","text":"Requirement already satisfied: onnx in /opt/conda/lib/python3.10/site-packages (1.14.0)\nRequirement already satisfied: numpy in /opt/conda/lib/python3.10/site-packages (from onnx) (1.23.5)\nRequirement already satisfied: protobuf>=3.20.2 in /opt/conda/lib/python3.10/site-packages (from onnx) (3.20.3)\nRequirement already satisfied: typing-extensions>=3.6.2.1 in /opt/conda/lib/python3.10/site-packages (from onnx) (4.5.0)\n\u001b[33mWARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv\u001b[0m\u001b[33m\n\u001b[0mCollecting onnx-simplifier\n  Downloading onnx_simplifier-0.4.33-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (2.2 MB)\n\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m2.2/2.2 MB\u001b[0m \u001b[31m23.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0ma \u001b[36m0:00:01\u001b[0m\n\u001b[?25hRequirement already satisfied: onnx in /opt/conda/lib/python3.10/site-packages (from onnx-simplifier) (1.14.0)\nRequirement already satisfied: rich in /opt/conda/lib/python3.10/site-packages (from onnx-simplifier) (12.6.0)\nRequirement already satisfied: numpy in /opt/conda/lib/python3.10/site-packages (from onnx->onnx-simplifier) (1.23.5)\nRequirement already satisfied: protobuf>=3.20.2 in /opt/conda/lib/python3.10/site-packages (from onnx->onnx-simplifier) (3.20.3)\nRequirement already satisfied: typing-extensions>=3.6.2.1 in /opt/conda/lib/python3.10/site-packages (from onnx->onnx-simplifier) (4.5.0)\nRequirement already satisfied: commonmark<0.10.0,>=0.9.0 in /opt/conda/lib/python3.10/site-packages (from rich->onnx-simplifier) (0.9.1)\nRequirement already satisfied: pygments<3.0.0,>=2.6.0 in /opt/conda/lib/python3.10/site-packages (from rich->onnx-simplifier) (2.15.1)\nInstalling collected packages: onnx-simplifier\nSuccessfully installed onnx-simplifier-0.4.33\n\u001b[33mWARNING: Running pip as the 'root' user can result in broken permissions and conflicting behaviour with the system package manager. It is recommended to use a virtual environment instead: https://pip.pypa.io/warnings/venv\u001b[0m\u001b[33m\n\u001b[0m","output_type":"stream"}]},{"cell_type":"code","source":"# from nn_meter import load_latency_predictor\nm_org = torch.load('/kaggle/input/pruning-effficientnet-v2-s-on-cifar10/best_model_9791.pt', map_location = torch.device('cpu')) #81% pruned model, unstructured\nm_pruned = torch.load('/kaggle/working/pruned_model_87%.pth',  map_location = torch.device('cpu')) #original model","metadata":{"execution":{"iopub.status.busy":"2023-07-25T01:35:41.649272Z","iopub.execute_input":"2023-07-25T01:35:41.649655Z","iopub.status.idle":"2023-07-25T01:35:43.145168Z","shell.execute_reply.started":"2023-07-25T01:35:41.649625Z","shell.execute_reply":"2023-07-25T01:35:43.143928Z"},"trusted":true},"execution_count":5,"outputs":[]},{"cell_type":"code","source":"from nn_meter import load_latency_predictor","metadata":{"execution":{"iopub.status.busy":"2023-07-25T01:35:51.448700Z","iopub.execute_input":"2023-07-25T01:35:51.449458Z","iopub.status.idle":"2023-07-25T01:35:52.438861Z","shell.execute_reply.started":"2023-07-25T01:35:51.449423Z","shell.execute_reply":"2023-07-25T01:35:52.437879Z"},"trusted":true},"execution_count":6,"outputs":[]},{"cell_type":"code","source":"predictor = load_latency_predictor('cortexA76cpu_tflite21', 1.0)\nlat_org = predictor.predict(m_org, 'torch')\nlat_pruned = predictor.predict(m_pruned, 'torch')","metadata":{"execution":{"iopub.status.busy":"2023-07-25T01:39:44.993395Z","iopub.execute_input":"2023-07-25T01:39:44.993773Z","iopub.status.idle":"2023-07-25T01:40:03.903951Z","shell.execute_reply.started":"2023-07-25T01:39:44.993732Z","shell.execute_reply":"2023-07-25T01:40:03.902966Z"},"trusted":true},"execution_count":13,"outputs":[{"name":"stdout","text":"(nn-Meter) checking local kernel predictors at /root/.nn_meter/data/predictor/cortexA76cpu_tflite21\n(nn-Meter) load predictor /root/.nn_meter/data/predictor/cortexA76cpu_tflite21/avgpool.pkl\n(nn-Meter) load predictor /root/.nn_meter/data/predictor/cortexA76cpu_tflite21/bn.pkl\n(nn-Meter) load predictor /root/.nn_meter/data/predictor/cortexA76cpu_tflite21/concat.pkl\n","output_type":"stream"},{"name":"stderr","text":"/opt/conda/lib/python3.10/site-packages/sklearn/base.py:318: UserWarning: Trying to unpickle estimator DecisionTreeRegressor from version 0.23.1 when using version 1.2.2. This might lead to breaking code or invalid results. Use at your own risk. For more info please refer to:\nhttps://scikit-learn.org/stable/model_persistence.html#security-maintainability-limitations\n  warnings.warn(\n/opt/conda/lib/python3.10/site-packages/sklearn/base.py:318: UserWarning: Trying to unpickle estimator RandomForestRegressor from version 0.23.1 when using version 1.2.2. This might lead to breaking code or invalid results. Use at your own risk. For more info please refer to:\nhttps://scikit-learn.org/stable/model_persistence.html#security-maintainability-limitations\n  warnings.warn(\n","output_type":"stream"},{"name":"stdout","text":"(nn-Meter) load predictor /root/.nn_meter/data/predictor/cortexA76cpu_tflite21/se.pkl\n(nn-Meter) load predictor /root/.nn_meter/data/predictor/cortexA76cpu_tflite21/dwconv-bn-relu.pkl\n(nn-Meter) load predictor /root/.nn_meter/data/predictor/cortexA76cpu_tflite21/relu.pkl\n(nn-Meter) load predictor /root/.nn_meter/data/predictor/cortexA76cpu_tflite21/bnrelu.pkl\n(nn-Meter) load predictor /root/.nn_meter/data/predictor/cortexA76cpu_tflite21/fc.pkl\n(nn-Meter) load predictor /root/.nn_meter/data/predictor/cortexA76cpu_tflite21/hswish.pkl\n(nn-Meter) load predictor /root/.nn_meter/data/predictor/cortexA76cpu_tflite21/add.pkl\n(nn-Meter) load predictor /root/.nn_meter/data/predictor/cortexA76cpu_tflite21/maxpool.pkl\n(nn-Meter) load predictor /root/.nn_meter/data/predictor/cortexA76cpu_tflite21/channelshuffle.pkl\n(nn-Meter) load predictor /root/.nn_meter/data/predictor/cortexA76cpu_tflite21/conv-bn-relu.pkl\n(nn-Meter) load predictor /root/.nn_meter/data/predictor/cortexA76cpu_tflite21/addrelu.pkl\n(nn-Meter) load predictor /root/.nn_meter/data/predictor/cortexA76cpu_tflite21/split.pkl\n(nn-Meter) load predictor /root/.nn_meter/data/predictor/cortexA76cpu_tflite21/global-avgpool.pkl\n(nn-Meter) Start latency prediction ...\n(nn-Meter) torch==2.0.0 is not well tested now, well tested version: torch==1.10.0, 1.9.0, 1.7.1\n(nn-Meter) Onnx-based Torch Converter is applied for model conversion\n(nn-Meter) onnx==1.14.0 is not well tested now, well tested version: onnx==1.10.0, 1.9.0\n(nn-Meter) torch==2.0.0 is not well tested now, well tested version: torch==1.10.0, 1.9.0, 1.7.1\n================ Diagnostic Run torch.onnx.export version 2.0.0 ================\nverbose: False, log level: Level.ERROR\n======================= 0 NONE 0 NOTE 0 WARNING 0 ERROR ========================\n\n(nn-Meter) onnx==1.14.0 is not well tested now, well tested version: onnx==1.10.0, 1.9.0\n(nn-Meter) Predict latency: 328.95630089686364 ms\n(nn-Meter) Start latency prediction ...\n(nn-Meter) torch==2.0.0 is not well tested now, well tested version: torch==1.10.0, 1.9.0, 1.7.1\n(nn-Meter) Onnx-based Torch Converter is applied for model conversion\n(nn-Meter) onnx==1.14.0 is not well tested now, well tested version: onnx==1.10.0, 1.9.0\n(nn-Meter) torch==2.0.0 is not well tested now, well tested version: torch==1.10.0, 1.9.0, 1.7.1\n================ Diagnostic Run torch.onnx.export version 2.0.0 ================\nverbose: False, log level: Level.ERROR\n======================= 0 NONE 0 NOTE 0 WARNING 0 ERROR ========================\n\n(nn-Meter) onnx==1.14.0 is not well tested now, well tested version: onnx==1.10.0, 1.9.0\n(nn-Meter) Predict latency: 128.62758512246805 ms\n","output_type":"stream"}]},{"cell_type":"code","source":"print(f'Latency of original model: {lat_org}')\nprint(f'Latency of pruned model: {lat_pruned}')","metadata":{"execution":{"iopub.status.busy":"2023-07-25T01:40:21.797561Z","iopub.execute_input":"2023-07-25T01:40:21.797999Z","iopub.status.idle":"2023-07-25T01:40:21.804961Z","shell.execute_reply.started":"2023-07-25T01:40:21.797965Z","shell.execute_reply":"2023-07-25T01:40:21.803506Z"},"trusted":true},"execution_count":14,"outputs":[{"name":"stdout","text":"Latency of original model: 328.95630089686364\nLatency of pruned model: 128.62758512246805\n","output_type":"stream"}]},{"cell_type":"code","source":"import os\nfile_size_bytes = os.path.getsize('/kaggle/input/pruning-effficientnet-v2-s-on-cifar10/best_model_9791.pt')\nfile_size_mb = file_size_bytes / (1024 * 1024)\nprint(f\"Model file size before pruning: {file_size_mb:.2f} MB\")\n\nfile_size_bytes = os.path.getsize('/kaggle/working/pruned_model_87%.pth')\nfile_size_mb = file_size_bytes / (1024 * 1024)\nprint(f\"Model file size after pruning: {file_size_mb:.2f} MB\")","metadata":{"execution":{"iopub.status.busy":"2023-07-25T01:54:55.995492Z","iopub.execute_input":"2023-07-25T01:54:55.995897Z","iopub.status.idle":"2023-07-25T01:54:56.003951Z","shell.execute_reply.started":"2023-07-25T01:54:55.995865Z","shell.execute_reply":"2023-07-25T01:54:56.002980Z"},"trusted":true},"execution_count":35,"outputs":[{"name":"stdout","text":"Model file size before pruning: 77.99 MB\nModel file size after pruning: 5.66 MB\n","output_type":"stream"}]},{"cell_type":"markdown","source":"# Ideas to generate insigths","metadata":{}},{"cell_type":"code","source":"#progressive retraining, channels are being pruned not parameters\n\n#ideas for plotting and generating insights, init size =80mb, accu=97.92\n#size and pruning \n#accuracy with prunning\n#mac redn vs pruned per vs prar dis\n#retrain epochs effect, recovery process\n#channel comparsion:intial and after pruned\n#Inference time comparison","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Replacing Silu with hswish and sigmoid with hsigmoid\nI need to deploy the model in FPG siLU is expensive","metadata":{}},{"cell_type":"code","source":"i=0\ndef replace_layer(m):\n    global i \n    for name, child in m.named_children():\n        if len(list(child.children()))> 0:\n            replace_layer(child)\n        \n        elif isinstance(child, nn.SiLU):\n            setattr(m, name, nn.Hardswish(inplace=True))\n            i=i+1\n\nreplace_layer(model)\nprint(f'{i} SiLU layers replaced with hard-swish layers')","metadata":{"execution":{"iopub.status.busy":"2023-07-25T01:43:12.670428Z","iopub.execute_input":"2023-07-25T01:43:12.670917Z","iopub.status.idle":"2023-07-25T01:43:12.693812Z","shell.execute_reply.started":"2023-07-25T01:43:12.670847Z","shell.execute_reply":"2023-07-25T01:43:12.692827Z"},"trusted":true},"execution_count":23,"outputs":[{"name":"stdout","text":"102 SiLU layers replaced with hard-swish layers\n","output_type":"stream"}]},{"cell_type":"code","source":"#checking if there is anu SiLU left\nfor module in model.modules():\n    if isinstance(module, nn.SiLU):\n        what=True\n    else: \n        what = False\n\nprint('SiLU Layers: ', what)","metadata":{"execution":{"iopub.status.busy":"2023-07-25T01:43:15.375359Z","iopub.execute_input":"2023-07-25T01:43:15.375733Z","iopub.status.idle":"2023-07-25T01:43:15.384775Z","shell.execute_reply.started":"2023-07-25T01:43:15.375703Z","shell.execute_reply":"2023-07-25T01:43:15.383560Z"},"trusted":true},"execution_count":24,"outputs":[{"name":"stdout","text":"SiLU Layers:  False\n","output_type":"stream"}]},{"cell_type":"code","source":"evaluate_model(model)","metadata":{"execution":{"iopub.status.busy":"2023-07-25T01:43:22.789757Z","iopub.execute_input":"2023-07-25T01:43:22.790492Z","iopub.status.idle":"2023-07-25T01:43:53.226108Z","shell.execute_reply.started":"2023-07-25T01:43:22.790458Z","shell.execute_reply":"2023-07-25T01:43:53.225067Z"},"trusted":true},"execution_count":26,"outputs":[{"name":"stderr","text":"100%|██████████| 313/313 [00:30<00:00, 10.30it/s]","output_type":"stream"},{"name":"stdout","text":"Evaluation Loss: 0.5765 Acc: 0.8418\n","output_type":"stream"},{"name":"stderr","text":"\n","output_type":"stream"},{"execution_count":26,"output_type":"execute_result","data":{"text/plain":"(0.576501197385788, tensor(0.8418, device='cuda:0', dtype=torch.float64))"},"metadata":{}}]},{"cell_type":"markdown","source":"some reduction in accuracy, Let's recover it by retraining it for few epochs","metadata":{}},{"cell_type":"code","source":"import numpy as np\nimport matplotlib.pyplot as plt\n\n# Define the sigmoid function\ndef sigmoid(x):\n    return 1 / (1 + np.exp(-x))\n\n# Define the hard sigmoid function\ndef hard_sigmoid(x):\n    return np.clip(x/6 + 0.5, 0, 1)\n\n# Generate data for the plot\nx = np.linspace(-10, 10, 1000)\ny_sigmoid = sigmoid(x)\ny_hard_sigmoid = hard_sigmoid(x)\n\n# Plot the sigmoid and hard sigmoid functions\nplt.figure(figsize=(10, 6))\nplt.plot(x, y_sigmoid, label='Sigmoid', color='blue')\nplt.plot(x, y_hard_sigmoid, label='Hard Sigmoid', color='red')\n\n# Set plot labels and title\nplt.xlabel('x')\nplt.ylabel('f(x)')\nplt.title('Sigmoid vs. Hard Sigmoid')\nplt.legend()\n\n# Show the plot\nplt.grid(True)\nplt.show()\n","metadata":{"execution":{"iopub.status.busy":"2023-07-25T01:49:26.603439Z","iopub.execute_input":"2023-07-25T01:49:26.603882Z","iopub.status.idle":"2023-07-25T01:49:27.014966Z","shell.execute_reply.started":"2023-07-25T01:49:26.603844Z","shell.execute_reply":"2023-07-25T01:49:27.013874Z"},"trusted":true},"execution_count":28,"outputs":[{"output_type":"display_data","data":{"text/plain":"<Figure size 1000x600 with 1 Axes>","image/png":"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"},"metadata":{}}]},{"cell_type":"code","source":"#need to replace sigmoid too\n\ni=0\ndef replace_layer(m):\n    global i \n    for name, child in m.named_children():\n        if len(list(child.children()))> 0:\n            replace_layer(child)\n        \n        elif isinstance(child, nn.Sigmoid):\n            setattr(m, name, nn.Hardsigmoid())\n            i=i+1\n\nreplace_layer(model)\nprint(f'{i} Sigmoid layers replaced with hard-sigmoid layers')\n","metadata":{"execution":{"iopub.status.busy":"2023-07-25T01:52:07.051441Z","iopub.execute_input":"2023-07-25T01:52:07.052237Z","iopub.status.idle":"2023-07-25T01:52:07.064630Z","shell.execute_reply.started":"2023-07-25T01:52:07.052200Z","shell.execute_reply":"2023-07-25T01:52:07.063409Z"},"trusted":true},"execution_count":32,"outputs":[{"name":"stdout","text":"30 Sigmoid layers replaced with hard-sigmoid layers\n","output_type":"stream"}]},{"cell_type":"code","source":"evaluate_model(model)","metadata":{"execution":{"iopub.status.busy":"2023-07-25T01:52:30.895345Z","iopub.execute_input":"2023-07-25T01:52:30.895740Z","iopub.status.idle":"2023-07-25T01:52:55.690901Z","shell.execute_reply.started":"2023-07-25T01:52:30.895709Z","shell.execute_reply":"2023-07-25T01:52:55.689884Z"},"trusted":true},"execution_count":33,"outputs":[{"name":"stderr","text":"100%|██████████| 313/313 [00:24<00:00, 12.63it/s]","output_type":"stream"},{"name":"stdout","text":"Evaluation Loss: 0.6638 Acc: 0.8217\n","output_type":"stream"},{"name":"stderr","text":"\n","output_type":"stream"},{"execution_count":33,"output_type":"execute_result","data":{"text/plain":"(0.6638343986749649, tensor(0.8217, device='cuda:0', dtype=torch.float64))"},"metadata":{}}]},{"cell_type":"markdown","source":"## Recovering the loss","metadata":{}},{"cell_type":"code","source":"# Further train for 20 more epochs\n\ncriterion = nn.CrossEntropyLoss()\n# Observe that all parameters are being optimized\noptimizer_ft = optim.SGD(model.parameters(), lr=0.001, momentum=0.9)\n\n# Decay LR by a factor of 0.1 every 7 epochs\nexp_lr_scheduler = lr_scheduler.StepLR(optimizer_ft, step_size=7, gamma=0.1)\n\nmodel, epoch_data = train_model(model, criterion, optimizer_ft, exp_lr_scheduler,\n                   num_epochs= 20 )\n #this is total pruned per\ntorch.save(model, f'EfficientnetV2_S_hswish87.pth')","metadata":{"execution":{"iopub.status.busy":"2023-07-25T01:55:58.445724Z","iopub.execute_input":"2023-07-25T01:55:58.446442Z","iopub.status.idle":"2023-07-25T03:27:44.122852Z","shell.execute_reply.started":"2023-07-25T01:55:58.446408Z","shell.execute_reply":"2023-07-25T03:27:44.121708Z"},"trusted":true},"execution_count":36,"outputs":[{"name":"stdout","text":"Epoch 1/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.1443 Acc: 0.9519\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.2258 Acc: 0.9286\n\nEpoch 2/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.1199 Acc: 0.9597\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.2040 Acc: 0.9367\n\nEpoch 3/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.1052 Acc: 0.9655\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.2075 Acc: 0.9357\n\nEpoch 4/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0930 Acc: 0.9692\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.2160 Acc: 0.9341\n\nEpoch 5/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0864 Acc: 0.9712\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.2067 Acc: 0.9367\n\nEpoch 6/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0811 Acc: 0.9729\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.2145 Acc: 0.9338\n\nEpoch 7/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0715 Acc: 0.9763\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.2023 Acc: 0.9383\n\nEpoch 8/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0494 Acc: 0.9849\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1847 Acc: 0.9456\n\nEpoch 9/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0393 Acc: 0.9880\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1831 Acc: 0.9462\n\nEpoch 10/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0348 Acc: 0.9894\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1834 Acc: 0.9456\n\nEpoch 11/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0320 Acc: 0.9913\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1825 Acc: 0.9478\n\nEpoch 12/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0310 Acc: 0.9912\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1839 Acc: 0.9472\n\nEpoch 13/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0293 Acc: 0.9916\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1836 Acc: 0.9480\n\nEpoch 14/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0279 Acc: 0.9919\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1861 Acc: 0.9478\n\nEpoch 15/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0261 Acc: 0.9926\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1844 Acc: 0.9488\n\nEpoch 16/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0241 Acc: 0.9934\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1866 Acc: 0.9480\n\nEpoch 17/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0241 Acc: 0.9935\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1835 Acc: 0.9481\n\nEpoch 18/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0252 Acc: 0.9930\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1848 Acc: 0.9480\n\nEpoch 19/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0261 Acc: 0.9925\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1849 Acc: 0.9483\n\nEpoch 20/20\n----------\n","output_type":"stream"},{"name":"stderr","text":"                                                   \r","output_type":"stream"},{"name":"stdout","text":"train Loss: 0.0246 Acc: 0.9933\n","output_type":"stream"},{"name":"stderr","text":"                                                 \r","output_type":"stream"},{"name":"stdout","text":"val Loss: 0.1869 Acc: 0.9478\n\nTraining complete in 91m 45s\nBest val Acc: 0.948800\n","output_type":"stream"}]}]}