Created
December 22, 2020 11:56
-
-
Save skbly7/88a69c509441dc8f92a30cafd8eeb8c0 to your computer and use it in GitHub Desktop.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| {"nbformat":4,"nbformat_minor":0,"metadata":{"colab":{"name":"Seismic Facies Identification Challenge BASELINE","provenance":[{"file_id":"1eCwQdXigALpJlz-zi2rZyyu39X-6_Oz8","timestamp":1601880140182}],"collapsed_sections":[]},"kernelspec":{"name":"python3","display_name":"Python 3"},"accelerator":"GPU","widgets":{"application/vnd.jupyter.widget-state+json":{"207a9f6bf37c4fa39c0ccac3669aaff1":{"model_module":"@jupyter-widgets/controls","model_name":"HBoxModel","state":{"_view_name":"HBoxView","_dom_classes":[],"_model_name":"HBoxModel","_view_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_view_count":null,"_view_module_version":"1.5.0","box_style":"","layout":"IPY_MODEL_e976c0fb58e54101bafdc1df2fd04b3d","_model_module":"@jupyter-widgets/controls","children":["IPY_MODEL_d71ed8197de44845a37ec2d55e68dccd","IPY_MODEL_496c9e4fa9a742f6b21c6ecda4e60499"]}},"e976c0fb58e54101bafdc1df2fd04b3d":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","state":{"_view_name":"LayoutView","grid_template_rows":null,"right":null,"justify_content":null,"_view_module":"@jupyter-widgets/base","overflow":null,"_model_module_version":"1.2.0","_view_count":null,"flex_flow":null,"width":null,"min_width":null,"border":null,"align_items":null,"bottom":null,"_model_module":"@jupyter-widgets/base","top":null,"grid_column":null,"overflow_y":null,"overflow_x":null,"grid_auto_flow":null,"grid_area":null,"grid_template_columns":null,"flex":null,"_model_name":"LayoutModel","justify_items":null,"grid_row":null,"max_height":null,"align_content":null,"visibility":null,"align_self":null,"height":null,"min_height":null,"padding":null,"grid_auto_rows":null,"grid_gap":null,"max_width":null,"order":null,"_view_module_version":"1.2.0","grid_template_areas":null,"object_position":null,"object_fit":null,"grid_auto_columns":null,"margin":null,"display":null,"left":null}},"d71ed8197de44845a37ec2d55e68dccd":{"model_module":"@jupyter-widgets/controls","model_name":"FloatProgressModel","state":{"_view_name":"ProgressView","style":"IPY_MODEL_3edb3bdb76e54a42b17cb1a7b8afd79e","_dom_classes":[],"description":"100%","_model_name":"FloatProgressModel","bar_style":"success","max":2012,"_view_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","value":2012,"_view_count":null,"_view_module_version":"1.5.0","orientation":"horizontal","min":0,"description_tooltip":null,"_model_module":"@jupyter-widgets/controls","layout":"IPY_MODEL_2f3ae18217ac4c96944917a7548f1578"}},"496c9e4fa9a742f6b21c6ecda4e60499":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","state":{"_view_name":"HTMLView","style":"IPY_MODEL_f37b6c64cb2e412194259df0002428a2","_dom_classes":[],"description":"","_model_name":"HTMLModel","placeholder":"","_view_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","value":" 2012/2012 [00:23<00:00, 87.03it/s]","_view_count":null,"_view_module_version":"1.5.0","description_tooltip":null,"_model_module":"@jupyter-widgets/controls","layout":"IPY_MODEL_7b38955330d8414097092e3f8724b827"}},"3edb3bdb76e54a42b17cb1a7b8afd79e":{"model_module":"@jupyter-widgets/controls","model_name":"ProgressStyleModel","state":{"_view_name":"StyleView","_model_name":"ProgressStyleModel","description_width":"initial","_view_module":"@jupyter-widgets/base","_model_module_version":"1.5.0","_view_count":null,"_view_module_version":"1.2.0","bar_color":null,"_model_module":"@jupyter-widgets/controls"}},"2f3ae18217ac4c96944917a7548f1578":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","state":{"_view_name":"LayoutView","grid_template_rows":null,"right":null,"justify_content":null,"_view_module":"@jupyter-widgets/base","overflow":null,"_model_module_version":"1.2.0","_view_count":null,"flex_flow":null,"width":null,"min_width":null,"border":null,"align_items":null,"bottom":null,"_model_module":"@jupyter-widgets/base","top":null,"grid_column":null,"overflow_y":null,"overflow_x":null,"grid_auto_flow":null,"grid_area":null,"grid_template_columns":null,"flex":null,"_model_name":"LayoutModel","justify_items":null,"grid_row":null,"max_height":null,"align_content":null,"visibility":null,"align_self":null,"height":null,"min_height":null,"padding":null,"grid_auto_rows":null,"grid_gap":null,"max_width":null,"order":null,"_view_module_version":"1.2.0","grid_template_areas":null,"object_position":null,"object_fit":null,"grid_auto_columns":null,"margin":null,"display":null,"left":null}},"f37b6c64cb2e412194259df0002428a2":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","state":{"_view_name":"StyleView","_model_name":"DescriptionStyleModel","description_width":"","_view_module":"@jupyter-widgets/base","_model_module_version":"1.5.0","_view_count":null,"_view_module_version":"1.2.0","_model_module":"@jupyter-widgets/controls"}},"7b38955330d8414097092e3f8724b827":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","state":{"_view_name":"LayoutView","grid_template_rows":null,"right":null,"justify_content":null,"_view_module":"@jupyter-widgets/base","overflow":null,"_model_module_version":"1.2.0","_view_count":null,"flex_flow":null,"width":null,"min_width":null,"border":null,"align_items":null,"bottom":null,"_model_module":"@jupyter-widgets/base","top":null,"grid_column":null,"overflow_y":null,"overflow_x":null,"grid_auto_flow":null,"grid_area":null,"grid_template_columns":null,"flex":null,"_model_name":"LayoutModel","justify_items":null,"grid_row":null,"max_height":null,"align_content":null,"visibility":null,"align_self":null,"height":null,"min_height":null,"padding":null,"grid_auto_rows":null,"grid_gap":null,"max_width":null,"order":null,"_view_module_version":"1.2.0","grid_template_areas":null,"object_position":null,"object_fit":null,"grid_auto_columns":null,"margin":null,"display":null,"left":null}},"4d86fb4b312641a3adec510933db7fbb":{"model_module":"@jupyter-widgets/controls","model_name":"HBoxModel","state":{"_view_name":"HBoxView","_dom_classes":[],"_model_name":"HBoxModel","_view_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","_view_count":null,"_view_module_version":"1.5.0","box_style":"","layout":"IPY_MODEL_e927fa08f72a4406a9eecdc3b3e01b77","_model_module":"@jupyter-widgets/controls","children":["IPY_MODEL_701d1708da484cd9b6dedc1dcf85a2ac","IPY_MODEL_edfc4455bfc242e7ab01550b4659c76a"]}},"e927fa08f72a4406a9eecdc3b3e01b77":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","state":{"_view_name":"LayoutView","grid_template_rows":null,"right":null,"justify_content":null,"_view_module":"@jupyter-widgets/base","overflow":null,"_model_module_version":"1.2.0","_view_count":null,"flex_flow":null,"width":null,"min_width":null,"border":null,"align_items":null,"bottom":null,"_model_module":"@jupyter-widgets/base","top":null,"grid_column":null,"overflow_y":null,"overflow_x":null,"grid_auto_flow":null,"grid_area":null,"grid_template_columns":null,"flex":null,"_model_name":"LayoutModel","justify_items":null,"grid_row":null,"max_height":null,"align_content":null,"visibility":null,"align_self":null,"height":null,"min_height":null,"padding":null,"grid_auto_rows":null,"grid_gap":null,"max_width":null,"order":null,"_view_module_version":"1.2.0","grid_template_areas":null,"object_position":null,"object_fit":null,"grid_auto_columns":null,"margin":null,"display":null,"left":null}},"701d1708da484cd9b6dedc1dcf85a2ac":{"model_module":"@jupyter-widgets/controls","model_name":"FloatProgressModel","state":{"_view_name":"ProgressView","style":"IPY_MODEL_f8fb31dd39c647269059b5f1201081ae","_dom_classes":[],"description":"100%","_model_name":"FloatProgressModel","bar_style":"success","max":1006,"_view_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","value":1006,"_view_count":null,"_view_module_version":"1.5.0","orientation":"horizontal","min":0,"description_tooltip":null,"_model_module":"@jupyter-widgets/controls","layout":"IPY_MODEL_18dc11f4daf642018da4e247ae97e958"}},"edfc4455bfc242e7ab01550b4659c76a":{"model_module":"@jupyter-widgets/controls","model_name":"HTMLModel","state":{"_view_name":"HTMLView","style":"IPY_MODEL_53f3daf48d934e9695f0a2065066fc04","_dom_classes":[],"description":"","_model_name":"HTMLModel","placeholder":"","_view_module":"@jupyter-widgets/controls","_model_module_version":"1.5.0","value":" 1006/1006 [00:02<00:00, 349.31it/s]","_view_count":null,"_view_module_version":"1.5.0","description_tooltip":null,"_model_module":"@jupyter-widgets/controls","layout":"IPY_MODEL_ba095a1eb12e475c90811383b5600c79"}},"f8fb31dd39c647269059b5f1201081ae":{"model_module":"@jupyter-widgets/controls","model_name":"ProgressStyleModel","state":{"_view_name":"StyleView","_model_name":"ProgressStyleModel","description_width":"initial","_view_module":"@jupyter-widgets/base","_model_module_version":"1.5.0","_view_count":null,"_view_module_version":"1.2.0","bar_color":null,"_model_module":"@jupyter-widgets/controls"}},"18dc11f4daf642018da4e247ae97e958":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","state":{"_view_name":"LayoutView","grid_template_rows":null,"right":null,"justify_content":null,"_view_module":"@jupyter-widgets/base","overflow":null,"_model_module_version":"1.2.0","_view_count":null,"flex_flow":null,"width":null,"min_width":null,"border":null,"align_items":null,"bottom":null,"_model_module":"@jupyter-widgets/base","top":null,"grid_column":null,"overflow_y":null,"overflow_x":null,"grid_auto_flow":null,"grid_area":null,"grid_template_columns":null,"flex":null,"_model_name":"LayoutModel","justify_items":null,"grid_row":null,"max_height":null,"align_content":null,"visibility":null,"align_self":null,"height":null,"min_height":null,"padding":null,"grid_auto_rows":null,"grid_gap":null,"max_width":null,"order":null,"_view_module_version":"1.2.0","grid_template_areas":null,"object_position":null,"object_fit":null,"grid_auto_columns":null,"margin":null,"display":null,"left":null}},"53f3daf48d934e9695f0a2065066fc04":{"model_module":"@jupyter-widgets/controls","model_name":"DescriptionStyleModel","state":{"_view_name":"StyleView","_model_name":"DescriptionStyleModel","description_width":"","_view_module":"@jupyter-widgets/base","_model_module_version":"1.5.0","_view_count":null,"_view_module_version":"1.2.0","_model_module":"@jupyter-widgets/controls"}},"ba095a1eb12e475c90811383b5600c79":{"model_module":"@jupyter-widgets/base","model_name":"LayoutModel","state":{"_view_name":"LayoutView","grid_template_rows":null,"right":null,"justify_content":null,"_view_module":"@jupyter-widgets/base","overflow":null,"_model_module_version":"1.2.0","_view_count":null,"flex_flow":null,"width":null,"min_width":null,"border":null,"align_items":null,"bottom":null,"_model_module":"@jupyter-widgets/base","top":null,"grid_column":null,"overflow_y":null,"overflow_x":null,"grid_auto_flow":null,"grid_area":null,"grid_template_columns":null,"flex":null,"_model_name":"LayoutModel","justify_items":null,"grid_row":null,"max_height":null,"align_content":null,"visibility":null,"align_self":null,"height":null,"min_height":null,"padding":null,"grid_auto_rows":null,"grid_gap":null,"max_width":null,"order":null,"_view_module_version":"1.2.0","grid_template_areas":null,"object_position":null,"object_fit":null,"grid_auto_columns":null,"margin":null,"display":null,"left":null}}}}},"cells":[{"cell_type":"markdown","metadata":{"id":"PMmb2MuQxJg-"},"source":["## Baseline Solution for [Seismic Facies Identification Challenge](https://www.aicrowd.com/challenges/seismic-facies-identification-challenge/leaderboards)"]},{"cell_type":"markdown","metadata":{"id":"b_u8z7CxxV-1"},"source":["### We used **U-Net** model whose architecture is shown below"]},{"cell_type":"markdown","metadata":{"id":"G3MOQ5-BuBU9"},"source":[""]},{"cell_type":"code","metadata":{"id":"Di_DYEnkIcFZ"},"source":["import torch.nn as nn\n","import torch\n","import torch.nn.functional as F\n","\n","# DoubleConv Class to perform two layer Convolution\n","class DoubleConv(nn.Module):\n"," def __init__(self,in_ch,out_ch):\n"," super(DoubleConv,self).__init__()\n"," self.conv = nn.Sequential(\n"," nn.Conv2d(in_ch,out_ch,3,padding=1), \n"," nn.BatchNorm2d(out_ch),\n"," nn.ReLU(inplace = True),\n"," nn.Conv2d(out_ch,out_ch,3,padding=1),\n"," nn.BatchNorm2d(out_ch),\n"," nn.ReLU(inplace = True)\n"," )\n"," def forward(self,x):\n"," return self.conv(x)\n","# Unter Class with shown architecture\n","class UNet(nn.Module):\n"," def __init__(self,in_ch,out_ch):\n"," super(UNet,self).__init__()\n"," self.conv1 = DoubleConv(in_ch,64)\n"," self.pool1 = nn.MaxPool2d(2)\n"," self.conv2 = DoubleConv(64,128)\n"," self.pool2 = nn.MaxPool2d(2)\n"," self.conv3 = DoubleConv(128,256)\n"," self.pool3 = nn.MaxPool2d(2)\n"," self.conv4 = DoubleConv(256,512)\n"," self.pool4 = nn.MaxPool2d(2)\n"," self.conv5 = DoubleConv(512,1024)\n"," self.up6 = nn.ConvTranspose2d(1024,512,2,stride=2)\n"," self.conv6 = DoubleConv(1024,512)\n"," self.up7 = nn.ConvTranspose2d(512,256,2,stride=2)\n"," self.conv7 = DoubleConv(512,256)\n"," self.up8 = nn.ConvTranspose2d(256,128,2,stride=2)\n"," self.conv8 = DoubleConv(256,128)\n"," self.up9 = nn.ConvTranspose2d(128,64,2,stride=2)\n"," self.conv9 = DoubleConv(128,128)\n"," self.conv9 = DoubleConv(128,64)\n"," self.conv10 = nn.Conv2d(64,out_ch,1)\n","\n"," def forward(self,x):\n"," c1 = self.conv1(x)\n"," p1 = self.pool1(c1)\n"," c2 = self.conv2(p1)\n"," p2 = self.pool2(c2)\n"," c3 = self.conv3(p2)\n"," p3 = self.pool3(c3)\n"," c4 = self.conv4(p3)\n"," p4 = self.pool4(c4)\n"," c5 = self.conv5(p4)\n"," up_6 = self.up6(c5)\n"," merge6 = torch.cat([up_6,c4],dim=1)\n"," c6 = self.conv6(merge6)\n"," up_7 = self.up7(c6)\n"," merge7 = torch.cat([up_7,c3],dim=1)\n"," c7 = self.conv7(merge7)\n"," up_8 = self.up8(c7)\n"," merge8 = torch.cat([up_8,c2],dim=1)\n"," c8 = self.conv8(merge8)\n"," up_9 = self.up9(c8)\n"," merge9 = torch.cat([up_9,c1],dim=1)\n"," c9 = self.conv9(merge9)\n"," out = c9\n"," c10 = self.conv10(c9)\n"," out = c10 \n"," out = nn.Softmax()(c10) \n"," return out"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"XiqISfeQx0U4"},"source":["### Data Download"]},{"cell_type":"code","metadata":{"id":"td-_AazWxyMZ"},"source":["# Getting the Data files from the aicrowd website\n","!wget https://datasets.aicrowd.com/default/aicrowd-public-datasets/seamai-facies-challenge/v0.1/public/data_train.npz\n","!wget https://datasets.aicrowd.com/default/aicrowd-public-datasets/seamai-facies-challenge/v0.1/public/data_test_1.npz\n","!wget https://datasets.aicrowd.com/default/aicrowd-public-datasets/seamai-facies-challenge/v0.1/public/labels_train.npz"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"ZzxwkPfZx_FQ"},"source":["#### If we look at the Train and the Test data we can observe that their shapes are different, So we decided to follow the Slice and Stack approach to make it nearly similar to the Test Data \n","#### (Look at the Below figure to understand)"]},{"cell_type":"markdown","metadata":{"id":"0kenoZxeybRr"},"source":[""]},{"cell_type":"markdown","metadata":{"id":"DN7zAkpw4oNl"},"source":["### Data Handling"]},{"cell_type":"code","metadata":{"id":"H9xKl5wrIkMg","executionInfo":{"status":"ok","timestamp":1601819481260,"user_tz":-330,"elapsed":50337,"user":{"displayName":"mani DILLS","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgX08ZK-rB1oDNubY9CHyhzUQ-aOYpijsySU8uQ=s64","userId":"00181444102423808384"}},"outputId":"1c4aac0c-871f-4f6b-d17b-905a7f8cc7b6","colab":{"base_uri":"https://localhost:8080/","height":34}},"source":["import time\n","import matplotlib.pyplot as plt\n","import numpy as np\n","import torch\n","from torch.utils.data import DataLoader, Dataset#\n","from torch import optim\n","\n","import numpy as np\n","sei_patch = np.load('/content/data_train.npz')['data']\n","lab_patch = np.load('/content/labels_train.npz')['labels']\n","\n","import cv2\n","from tqdm.notebook import tqdm\n","import datetime \n","from IPython.display import HTML\n","import cv2\n","\n","\n","sei_patch = np.concatenate((sei_patch[:,:,:295] , sei_patch[:,:,295:]))\n","lab_patch = np.concatenate((lab_patch[:,:,:295] , lab_patch[:,:,295:]))\n","sei_patch.shape, lab_patch.shape"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["((2012, 782, 295), (2012, 782, 295))"]},"metadata":{"tags":[]},"execution_count":2}]},{"cell_type":"markdown","metadata":{"id":"coVT7uMi2xg4"},"source":["### Resizing the images Considering the Computational power"]},{"cell_type":"code","metadata":{"id":"TO_labAzItyD","executionInfo":{"status":"ok","timestamp":1601819503686,"user_tz":-330,"elapsed":9265,"user":{"displayName":"mani DILLS","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgX08ZK-rB1oDNubY9CHyhzUQ-aOYpijsySU8uQ=s64","userId":"00181444102423808384"}},"outputId":"e48c619c-b400-428d-eb5e-8dfeb7019f14","colab":{"base_uri":"https://localhost:8080/","height":84,"referenced_widgets":["207a9f6bf37c4fa39c0ccac3669aaff1","e976c0fb58e54101bafdc1df2fd04b3d","d71ed8197de44845a37ec2d55e68dccd","496c9e4fa9a742f6b21c6ecda4e60499","3edb3bdb76e54a42b17cb1a7b8afd79e","2f3ae18217ac4c96944917a7548f1578","f37b6c64cb2e412194259df0002428a2","7b38955330d8414097092e3f8724b827"]}},"source":["training_img_data = []\n","training_label_data = []\n","\n","for i in tqdm(range(0, 2012)):\n"," img = sei_patch[i,:,:]\n"," label = lab_patch[i,:,:]\n","\n"," img = np.expand_dims(img, axis=2).astype('float32')\n"," label = np.expand_dims(label, axis=2).astype('float32')\n","\n"," img = cv2.resize(img, (256,256 ))\n"," label = cv2.resize(label, ( 256,256))\n","\n"," training_img_data.append(img) \n"," training_label_data.append(label)\n","\n","training_img_data = np.asarray(training_img_data)\n","training_label_data = np.asarray(training_label_data)\n","training_label_data = np.array(training_label_data,dtype=int)\n","training_img_data.shape, training_label_data.shape"],"execution_count":null,"outputs":[{"output_type":"display_data","data":{"application/vnd.jupyter.widget-view+json":{"model_id":"207a9f6bf37c4fa39c0ccac3669aaff1","version_minor":0,"version_major":2},"text/plain":["HBox(children=(FloatProgress(value=0.0, max=2012.0), HTML(value='')))"]},"metadata":{"tags":[]}},{"output_type":"stream","text":["\n"],"name":"stdout"},{"output_type":"execute_result","data":{"text/plain":["((2012, 256, 256), (2012, 256, 256))"]},"metadata":{"tags":[]},"execution_count":3}]},{"cell_type":"markdown","metadata":{"id":"DhuSx5sK3iuo"},"source":["### Creating a DataGenerator Class along with Metrics and Train Function"]},{"cell_type":"code","metadata":{"id":"UYDYvdV6I9sP"},"source":["class DataGenerator(Dataset):\n"," def __init__(self, x_set, y_set):\n"," self.x, self.y = x_set, y_set\n","\n"," def __len__(self):\n"," return len(self.x)\n"," \n"," def __getitem__(self, index):\n"," batch_x = self.x[index]\n"," batch_y = self.y[index]\n"," return np.expand_dims(batch_x,axis=0), batch_y\n","\n","e=1e-2\n","\n","def accuracy(out, yb):\n"," preds = torch.argmax(out, dim=1)\n"," return (preds == yb).float().mean()\n","\n","\n","def train(model,optimizer,dataload,num_epochs,device):\n"," acc_history = []\n"," loss_history = []\n"," miou_history = []\n"," for epoch in range(num_epochs):\n"," print('Starting epoch {}/{}'.format(epoch+1, num_epochs))\n"," print('-' * 10)\n"," since = time.time()\n"," dataset_size = len(dataload.dataset)\n"," epoch_loss = 0\n"," epoch_acc = 0\n","\n"," for idx,(x, y) in enumerate(dataload): \n"," optimizer.zero_grad() \n"," inputs = x.to(device)\n"," labels = y.to(device)\n"," outputs = model(inputs) \n"," criterion1 = MulticlassDiceLoss() \n"," loss1 = criterion1(outputs,labels.long())\n"," criterion2 = torch.nn.CrossEntropyLoss()\n"," loss2 = criterion2(torch.log(outputs),labels.long())\n"," loss = e*loss1+loss2\n"," acc = accuracy(outputs,labels)\n"," loss.backward() \n"," optimizer.step() \n"," \n"," epoch_loss += loss.item()\n"," epoch_acc+= acc\n"," loss_history.append(loss.item())\n"," acc_history.append(acc)\n"," if (idx+1)%10==0:\n"," print(\"%d/%d,train_loss:%0.3f,accuracy:%0.3f\" % (idx+1, dataset_size // dataload.batch_size, loss.item(),acc))\n","\n"," time_elapsed = time.time() - since \n"," all_epoch_loss=epoch_loss/len(dataload)\n"," all_epoch_acc=epoch_acc/len(dataload)\n"," print(\"epoch %d loss:%0.3f accuracy:%0.3f \" % (epoch, all_epoch_loss,all_epoch_acc))\n"," print('Training complete in {:.0f}m {:.0f}s'.format(time_elapsed // 60, time_elapsed % 60))\n","\n"," torch.save(model,\"/content/model_0.pth\") \n"," return model,loss_history,acc_history"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"HQfWz9Fw3b1N"},"source":["### Data Loading and Model Training"]},{"cell_type":"code","metadata":{"id":"YOXQFVITJH3Z","executionInfo":{"status":"ok","timestamp":1601820899801,"user_tz":-330,"elapsed":1389944,"user":{"displayName":"mani DILLS","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgX08ZK-rB1oDNubY9CHyhzUQ-aOYpijsySU8uQ=s64","userId":"00181444102423808384"}},"outputId":"35364277-63ad-4603-f572-fe4657b6e2f6","colab":{"base_uri":"https://localhost:8080/","height":1000}},"source":["from dice_loss import *\n","!CUDA_LAUNCH_BLOCKING=1\n","device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n","model = UNet(1,7).to(device)\n","train_dataset = DataGenerator(x_set=training_img_data,y_set=training_label_data) \n","dataloader = DataLoader(train_dataset, batch_size=10, shuffle=True)\n","optimizer = optim.Adam(model.parameters(),lr=1e-4)\n","num_epochs=12\n","\n","model_0,loss,acc=train(model,optimizer,dataloader,num_epochs,device)"],"execution_count":null,"outputs":[{"output_type":"stream","text":["Starting epoch 1/12\n","----------\n"],"name":"stdout"},{"output_type":"stream","text":["/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:67: UserWarning: Implicit dimension choice for softmax has been deprecated. Change the call to include dim=X as an argument.\n"],"name":"stderr"},{"output_type":"stream","text":["10/201,train_loss:2.073,accuracy:0.211\n","20/201,train_loss:1.849,accuracy:0.339\n","30/201,train_loss:1.520,accuracy:0.538\n","40/201,train_loss:1.539,accuracy:0.506\n","50/201,train_loss:1.597,accuracy:0.461\n","60/201,train_loss:1.550,accuracy:0.508\n","70/201,train_loss:1.272,accuracy:0.652\n","80/201,train_loss:1.069,accuracy:0.780\n","90/201,train_loss:1.375,accuracy:0.585\n","100/201,train_loss:1.562,accuracy:0.537\n","110/201,train_loss:1.281,accuracy:0.626\n","120/201,train_loss:1.114,accuracy:0.743\n","130/201,train_loss:1.379,accuracy:0.597\n","140/201,train_loss:1.265,accuracy:0.555\n","150/201,train_loss:1.633,accuracy:0.541\n","160/201,train_loss:1.098,accuracy:0.694\n","170/201,train_loss:1.064,accuracy:0.709\n","180/201,train_loss:1.111,accuracy:0.678\n","190/201,train_loss:0.999,accuracy:0.734\n","200/201,train_loss:1.140,accuracy:0.678\n","epoch 0 loss:1.421 accuracy:0.567 \n","Training complete in 1m 53s\n","Starting epoch 2/12\n","----------\n","10/201,train_loss:1.276,accuracy:0.576\n","20/201,train_loss:1.067,accuracy:0.680\n","30/201,train_loss:1.597,accuracy:0.460\n","40/201,train_loss:1.233,accuracy:0.585\n","50/201,train_loss:1.260,accuracy:0.575\n","60/201,train_loss:0.845,accuracy:0.794\n","70/201,train_loss:1.002,accuracy:0.675\n","80/201,train_loss:0.905,accuracy:0.775\n","90/201,train_loss:0.880,accuracy:0.776\n","100/201,train_loss:0.843,accuracy:0.780\n","110/201,train_loss:1.156,accuracy:0.659\n","120/201,train_loss:0.919,accuracy:0.737\n","130/201,train_loss:0.702,accuracy:0.845\n","140/201,train_loss:1.032,accuracy:0.686\n","150/201,train_loss:0.779,accuracy:0.804\n","160/201,train_loss:0.808,accuracy:0.780\n","170/201,train_loss:0.766,accuracy:0.817\n","180/201,train_loss:0.608,accuracy:0.857\n","190/201,train_loss:0.834,accuracy:0.752\n","200/201,train_loss:0.738,accuracy:0.779\n","epoch 1 loss:0.993 accuracy:0.712 \n","Training complete in 1m 55s\n","Starting epoch 3/12\n","----------\n","10/201,train_loss:1.025,accuracy:0.682\n","20/201,train_loss:0.727,accuracy:0.827\n","30/201,train_loss:1.059,accuracy:0.653\n","40/201,train_loss:0.771,accuracy:0.762\n","50/201,train_loss:0.869,accuracy:0.731\n","60/201,train_loss:0.684,accuracy:0.816\n","70/201,train_loss:0.936,accuracy:0.745\n","80/201,train_loss:0.543,accuracy:0.891\n","90/201,train_loss:0.758,accuracy:0.788\n","100/201,train_loss:1.130,accuracy:0.654\n","110/201,train_loss:0.679,accuracy:0.836\n","120/201,train_loss:0.580,accuracy:0.854\n","130/201,train_loss:0.911,accuracy:0.719\n","140/201,train_loss:0.858,accuracy:0.777\n","150/201,train_loss:0.475,accuracy:0.913\n","160/201,train_loss:0.510,accuracy:0.890\n","170/201,train_loss:0.472,accuracy:0.905\n","180/201,train_loss:0.755,accuracy:0.770\n","190/201,train_loss:0.536,accuracy:0.877\n","200/201,train_loss:0.502,accuracy:0.884\n","epoch 2 loss:0.761 accuracy:0.790 \n","Training complete in 1m 55s\n","Starting epoch 4/12\n","----------\n","10/201,train_loss:0.767,accuracy:0.770\n","20/201,train_loss:0.647,accuracy:0.821\n","30/201,train_loss:0.664,accuracy:0.812\n","40/201,train_loss:0.688,accuracy:0.818\n","50/201,train_loss:0.641,accuracy:0.800\n","60/201,train_loss:0.889,accuracy:0.692\n","70/201,train_loss:0.458,accuracy:0.906\n","80/201,train_loss:0.635,accuracy:0.827\n","90/201,train_loss:0.779,accuracy:0.784\n","100/201,train_loss:0.708,accuracy:0.763\n","110/201,train_loss:0.421,accuracy:0.920\n","120/201,train_loss:0.582,accuracy:0.858\n","130/201,train_loss:0.639,accuracy:0.822\n","140/201,train_loss:0.400,accuracy:0.918\n","150/201,train_loss:0.513,accuracy:0.862\n","160/201,train_loss:0.487,accuracy:0.877\n","170/201,train_loss:0.490,accuracy:0.891\n","180/201,train_loss:0.578,accuracy:0.848\n","190/201,train_loss:0.657,accuracy:0.790\n","200/201,train_loss:0.793,accuracy:0.764\n","epoch 3 loss:0.587 accuracy:0.848 \n","Training complete in 1m 55s\n","Starting epoch 5/12\n","----------\n","10/201,train_loss:0.440,accuracy:0.911\n","20/201,train_loss:0.635,accuracy:0.811\n","30/201,train_loss:0.416,accuracy:0.903\n","40/201,train_loss:0.433,accuracy:0.890\n","50/201,train_loss:0.407,accuracy:0.915\n","60/201,train_loss:0.402,accuracy:0.932\n","70/201,train_loss:0.678,accuracy:0.841\n","80/201,train_loss:0.370,accuracy:0.929\n","90/201,train_loss:0.458,accuracy:0.883\n","100/201,train_loss:0.378,accuracy:0.932\n","110/201,train_loss:0.563,accuracy:0.850\n","120/201,train_loss:0.363,accuracy:0.919\n","130/201,train_loss:0.328,accuracy:0.944\n","140/201,train_loss:0.462,accuracy:0.881\n","150/201,train_loss:0.335,accuracy:0.936\n","160/201,train_loss:0.276,accuracy:0.959\n","170/201,train_loss:0.395,accuracy:0.911\n","180/201,train_loss:0.289,accuracy:0.952\n","190/201,train_loss:0.370,accuracy:0.914\n","200/201,train_loss:0.532,accuracy:0.860\n","epoch 4 loss:0.446 accuracy:0.896 \n","Training complete in 1m 55s\n","Starting epoch 6/12\n","----------\n","10/201,train_loss:0.611,accuracy:0.822\n","20/201,train_loss:0.307,accuracy:0.939\n","30/201,train_loss:0.598,accuracy:0.876\n","40/201,train_loss:0.329,accuracy:0.928\n","50/201,train_loss:0.380,accuracy:0.909\n","60/201,train_loss:0.336,accuracy:0.929\n","70/201,train_loss:0.455,accuracy:0.888\n","80/201,train_loss:0.369,accuracy:0.908\n","90/201,train_loss:0.455,accuracy:0.871\n","100/201,train_loss:0.343,accuracy:0.912\n","110/201,train_loss:0.319,accuracy:0.940\n","120/201,train_loss:0.344,accuracy:0.924\n","130/201,train_loss:0.249,accuracy:0.959\n","140/201,train_loss:0.323,accuracy:0.934\n","150/201,train_loss:0.312,accuracy:0.949\n","160/201,train_loss:0.270,accuracy:0.949\n","170/201,train_loss:0.275,accuracy:0.949\n","180/201,train_loss:0.326,accuracy:0.934\n","190/201,train_loss:0.439,accuracy:0.890\n","200/201,train_loss:0.428,accuracy:0.900\n","epoch 5 loss:0.383 accuracy:0.913 \n","Training complete in 1m 55s\n","Starting epoch 7/12\n","----------\n","10/201,train_loss:0.470,accuracy:0.862\n","20/201,train_loss:0.299,accuracy:0.944\n","30/201,train_loss:0.274,accuracy:0.948\n","40/201,train_loss:0.293,accuracy:0.950\n","50/201,train_loss:1.170,accuracy:0.747\n","60/201,train_loss:0.304,accuracy:0.940\n","70/201,train_loss:0.513,accuracy:0.883\n","80/201,train_loss:0.407,accuracy:0.894\n","90/201,train_loss:0.312,accuracy:0.942\n","100/201,train_loss:0.213,accuracy:0.969\n","110/201,train_loss:0.259,accuracy:0.954\n","120/201,train_loss:0.333,accuracy:0.929\n","130/201,train_loss:0.301,accuracy:0.939\n","140/201,train_loss:0.341,accuracy:0.922\n","150/201,train_loss:0.240,accuracy:0.963\n","160/201,train_loss:0.223,accuracy:0.969\n","170/201,train_loss:0.250,accuracy:0.956\n","180/201,train_loss:0.237,accuracy:0.958\n","190/201,train_loss:0.312,accuracy:0.942\n","200/201,train_loss:0.256,accuracy:0.949\n","epoch 6 loss:0.337 accuracy:0.927 \n","Training complete in 1m 55s\n","Starting epoch 8/12\n","----------\n","10/201,train_loss:0.235,accuracy:0.961\n","20/201,train_loss:0.266,accuracy:0.945\n","30/201,train_loss:0.192,accuracy:0.975\n","40/201,train_loss:0.239,accuracy:0.955\n","50/201,train_loss:0.272,accuracy:0.934\n","60/201,train_loss:0.230,accuracy:0.960\n","70/201,train_loss:0.272,accuracy:0.948\n","80/201,train_loss:0.271,accuracy:0.948\n","90/201,train_loss:0.245,accuracy:0.949\n","100/201,train_loss:0.261,accuracy:0.948\n","110/201,train_loss:0.243,accuracy:0.953\n","120/201,train_loss:0.233,accuracy:0.957\n","130/201,train_loss:0.187,accuracy:0.972\n","140/201,train_loss:0.283,accuracy:0.934\n","150/201,train_loss:0.211,accuracy:0.975\n","160/201,train_loss:0.208,accuracy:0.963\n","170/201,train_loss:0.230,accuracy:0.965\n","180/201,train_loss:0.249,accuracy:0.946\n","190/201,train_loss:0.271,accuracy:0.945\n","200/201,train_loss:0.254,accuracy:0.946\n","epoch 7 loss:0.262 accuracy:0.949 \n","Training complete in 1m 55s\n","Starting epoch 9/12\n","----------\n","10/201,train_loss:0.208,accuracy:0.965\n","20/201,train_loss:0.968,accuracy:0.779\n","30/201,train_loss:0.303,accuracy:0.916\n","40/201,train_loss:0.253,accuracy:0.954\n","50/201,train_loss:0.309,accuracy:0.927\n","60/201,train_loss:0.294,accuracy:0.935\n","70/201,train_loss:0.342,accuracy:0.919\n","80/201,train_loss:0.334,accuracy:0.920\n","90/201,train_loss:0.246,accuracy:0.956\n","100/201,train_loss:0.346,accuracy:0.903\n","110/201,train_loss:0.293,accuracy:0.936\n","120/201,train_loss:0.224,accuracy:0.957\n","130/201,train_loss:0.246,accuracy:0.956\n","140/201,train_loss:0.232,accuracy:0.959\n","150/201,train_loss:0.192,accuracy:0.965\n","160/201,train_loss:0.210,accuracy:0.965\n","170/201,train_loss:0.220,accuracy:0.960\n","180/201,train_loss:0.190,accuracy:0.968\n","190/201,train_loss:0.268,accuracy:0.952\n","200/201,train_loss:0.190,accuracy:0.970\n","epoch 8 loss:0.290 accuracy:0.936 \n","Training complete in 1m 55s\n","Starting epoch 10/12\n","----------\n","10/201,train_loss:0.247,accuracy:0.946\n","20/201,train_loss:0.186,accuracy:0.971\n","30/201,train_loss:0.243,accuracy:0.955\n","40/201,train_loss:0.347,accuracy:0.919\n","50/201,train_loss:0.172,accuracy:0.977\n","60/201,train_loss:0.173,accuracy:0.975\n","70/201,train_loss:0.193,accuracy:0.970\n","80/201,train_loss:0.159,accuracy:0.979\n","90/201,train_loss:0.225,accuracy:0.950\n","100/201,train_loss:0.177,accuracy:0.969\n","110/201,train_loss:0.156,accuracy:0.975\n","120/201,train_loss:0.173,accuracy:0.973\n","130/201,train_loss:0.164,accuracy:0.974\n","140/201,train_loss:0.166,accuracy:0.978\n","150/201,train_loss:0.184,accuracy:0.968\n","160/201,train_loss:0.167,accuracy:0.969\n","170/201,train_loss:0.188,accuracy:0.972\n","180/201,train_loss:0.421,accuracy:0.900\n","190/201,train_loss:0.228,accuracy:0.951\n","200/201,train_loss:0.203,accuracy:0.961\n","epoch 9 loss:0.218 accuracy:0.959 \n","Training complete in 1m 55s\n","Starting epoch 11/12\n","----------\n","10/201,train_loss:0.258,accuracy:0.938\n","20/201,train_loss:0.186,accuracy:0.965\n","30/201,train_loss:0.241,accuracy:0.947\n","40/201,train_loss:0.195,accuracy:0.965\n","50/201,train_loss:0.157,accuracy:0.976\n","60/201,train_loss:0.216,accuracy:0.954\n","70/201,train_loss:0.321,accuracy:0.919\n","80/201,train_loss:0.233,accuracy:0.942\n","90/201,train_loss:0.150,accuracy:0.979\n","100/201,train_loss:0.181,accuracy:0.967\n","110/201,train_loss:0.148,accuracy:0.979\n","120/201,train_loss:0.173,accuracy:0.968\n","130/201,train_loss:0.211,accuracy:0.967\n","140/201,train_loss:0.176,accuracy:0.966\n","150/201,train_loss:0.198,accuracy:0.963\n","160/201,train_loss:0.190,accuracy:0.966\n","170/201,train_loss:0.163,accuracy:0.971\n","180/201,train_loss:0.195,accuracy:0.965\n","190/201,train_loss:0.186,accuracy:0.964\n","200/201,train_loss:0.198,accuracy:0.961\n","epoch 10 loss:0.209 accuracy:0.959 \n","Training complete in 1m 55s\n","Starting epoch 12/12\n","----------\n","10/201,train_loss:0.317,accuracy:0.918\n","20/201,train_loss:0.322,accuracy:0.926\n","30/201,train_loss:0.194,accuracy:0.967\n","40/201,train_loss:0.208,accuracy:0.962\n","50/201,train_loss:0.175,accuracy:0.960\n","60/201,train_loss:0.179,accuracy:0.965\n","70/201,train_loss:0.219,accuracy:0.967\n","80/201,train_loss:0.150,accuracy:0.978\n","90/201,train_loss:0.176,accuracy:0.966\n","100/201,train_loss:0.216,accuracy:0.953\n","110/201,train_loss:0.165,accuracy:0.972\n","120/201,train_loss:0.187,accuracy:0.976\n","130/201,train_loss:0.165,accuracy:0.977\n","140/201,train_loss:0.223,accuracy:0.960\n","150/201,train_loss:0.142,accuracy:0.979\n","160/201,train_loss:0.148,accuracy:0.981\n","170/201,train_loss:0.209,accuracy:0.961\n","180/201,train_loss:0.146,accuracy:0.979\n","190/201,train_loss:0.188,accuracy:0.963\n","200/201,train_loss:0.145,accuracy:0.979\n","epoch 11 loss:0.178 accuracy:0.968 \n","Training complete in 1m 55s\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"4WxI5bwAJQG3"},"source":["## Function to predict\n","def seisfacies_predict(section,patch_size=256,overlap=0,onehot=0): \n"," m1,m2 = section.shape\n"," os = overlap \n"," n1,n2 = patch_size,patch_size \n"," c1 = int(np.round((m1+os)/(n1-os)+0.5))\n"," c2 = int(np.round((m2+os)/(n2-os)+0.5))\n"," p1 = (n1-os)*c1+os\n"," p2 = (n2-os)*c2+os\n","\n"," gp = np.zeros((p1,p2),dtype=np.single) \n"," gy = np.zeros((7,p1,p2),dtype=np.single) \n"," gs = np.zeros((n1,n2),dtype=np.single) \n"," \n"," gp[0:m1,0:m2]=section \n","\n"," for k1 in range(c1):\n"," for k2 in range(c2):\n"," b1 = k1*n1-k1*os\n"," e1 = b1+n1\n"," b2 = k2*n2-k2*os\n"," e2 = b2+n2 \n"," #predict\n"," gs[:,:]=gp[b1:e1,b2:e2]\n"," x=gs.reshape(1,1,256,256)\n"," Y_patch= model(torch.from_numpy(x)).squeeze()\n"," p=F.softmax(Y_patch, dim=0).detach().numpy()\n"," gy[:,b1:e1,b2:e2]= gy[:,b1:e1,b2:e2]+p\n"," \n"," gy_onehot = gy[:,0:m1,0:m2] \n"," #onehot2label\n"," gy_label =np.argmax(gy_onehot,axis=0)\n","\n"," if onehot==0:\n"," return gy_label\n"," if onehot==1:\n"," return gy_label,gy_onehot"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"hn0bpfnC4S_4"},"source":["### Loading Test Data"]},{"cell_type":"code","metadata":{"id":"0JPlcswMMucV","executionInfo":{"status":"ok","timestamp":1601663189780,"user_tz":-330,"elapsed":24178,"user":{"displayName":"Mani Kandan","photoUrl":"","userId":"16019920751672798074"}},"outputId":"4d92636f-da70-4db9-ccc9-585ce77dc01d","colab":{"base_uri":"https://localhost:8080/","height":66,"referenced_widgets":["4d86fb4b312641a3adec510933db7fbb","e927fa08f72a4406a9eecdc3b3e01b77","701d1708da484cd9b6dedc1dcf85a2ac","edfc4455bfc242e7ab01550b4659c76a","f8fb31dd39c647269059b5f1201081ae","18dc11f4daf642018da4e247ae97e958","53f3daf48d934e9695f0a2065066fc04","ba095a1eb12e475c90811383b5600c79"]}},"source":["test_seismic = np.load('/content/data_test_1.npz')['data']\n","testing_img_data = []\n","for i in tqdm(range(0, 1006)):\n"," img = test_seismic[i,:,:]\n"," img = np.expand_dims(img, axis=2).astype('float32')\n"," img = cv2.resize(img, (256, 256))\n"," testing_img_data.append(img) \n"," \n","testing_img_data = np.asarray(testing_img_data)\n"],"execution_count":null,"outputs":[{"output_type":"display_data","data":{"application/vnd.jupyter.widget-view+json":{"model_id":"4d86fb4b312641a3adec510933db7fbb","version_minor":0,"version_major":2},"text/plain":["HBox(children=(FloatProgress(value=0.0, max=1006.0), HTML(value='')))"]},"metadata":{"tags":[]}},{"output_type":"stream","text":["\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"SVZLZbiA4A-e"},"source":["### Testing Predictions for One Data"]},{"cell_type":"code","metadata":{"id":"NuogSaK8MxGi","executionInfo":{"status":"ok","timestamp":1601821115530,"user_tz":-330,"elapsed":7374,"user":{"displayName":"mani DILLS","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgX08ZK-rB1oDNubY9CHyhzUQ-aOYpijsySU8uQ=s64","userId":"00181444102423808384"}},"outputId":"0823b072-fd8c-47b9-f0c5-1a638785df10","colab":{"base_uri":"https://localhost:8080/","height":176}},"source":["#loading the Saved Model\n","model = torch.load(\"model_0.pth\",map_location='cpu')\n","gy_label,gy_onehot=seisfacies_predict(training_img_data[180],onehot=1)\n","gy_label"],"execution_count":null,"outputs":[{"output_type":"stream","text":["/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:67: UserWarning: Implicit dimension choice for softmax has been deprecated. Change the call to include dim=X as an argument.\n"],"name":"stderr"},{"output_type":"execute_result","data":{"text/plain":["array([[2, 2, 2, ..., 2, 2, 2],\n"," [2, 2, 2, ..., 2, 2, 2],\n"," [2, 6, 6, ..., 2, 2, 2],\n"," ...,\n"," [6, 6, 6, ..., 4, 4, 4],\n"," [6, 6, 6, ..., 4, 4, 4],\n"," [2, 6, 6, ..., 4, 4, 4]])"]},"metadata":{"tags":[]},"execution_count":14}]},{"cell_type":"code","metadata":{"id":"AU7bIlHsM4gQ","executionInfo":{"status":"ok","timestamp":1601821120525,"user_tz":-330,"elapsed":1172,"user":{"displayName":"mani DILLS","photoUrl":"https://lh3.googleusercontent.com/a-/AOh14GgX08ZK-rB1oDNubY9CHyhzUQ-aOYpijsySU8uQ=s64","userId":"00181444102423808384"}},"outputId":"e470bafc-1029-4a6c-ef7b-5d5d06159007","colab":{"base_uri":"https://localhost:8080/","height":139}},"source":["training_label_data[180]"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([[6, 6, 6, ..., 2, 2, 2],\n"," [6, 6, 6, ..., 2, 2, 2],\n"," [6, 6, 6, ..., 2, 2, 2],\n"," ...,\n"," [6, 6, 6, ..., 4, 4, 4],\n"," [6, 6, 6, ..., 4, 4, 4],\n"," [6, 6, 6, ..., 4, 4, 4]])"]},"metadata":{"tags":[]},"execution_count":15}]},{"cell_type":"markdown","metadata":{"id":"9zBBTnLW39xO"},"source":["### Predictions"]},{"cell_type":"code","metadata":{"id":"OqRSmjtQM5jQ"},"source":["preds = []\n","for i in tqdm(range(0, 1006)):\n"," gy_label,gy_onehot=seisfacies_predict(testing_img_data[i],onehot=1)\n"," label = gy_label\n"," label = np.expand_dims(label, axis=2).astype('float32')\n"," label = cv2.resize(label, (251,782))\n"," preds.append(label)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"1-xy2j2H2Oy8"},"source":["final_preds = np.asarray(preds)\n","final_preds = np.array(final_preds,dtype=int)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"hpsSCWbl2R0j"},"source":["np.savez_compressed('submission.npz', prediction = final_preds)"],"execution_count":null,"outputs":[]}]} |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment