{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.10.12","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"kaggle":{"accelerator":"none","dataSources":[{"sourceId":52254,"databundleVersionId":6863140,"sourceType":"competition"},{"sourceId":6523471,"sourceType":"datasetVersion","datasetId":3771357},{"sourceId":6524344,"sourceType":"datasetVersion","datasetId":3771912},{"sourceId":7432254,"sourceType":"datasetVersion","datasetId":4325089},{"sourceId":7665407,"sourceType":"datasetVersion","datasetId":4470305}],"dockerImageVersionId":30626,"isInternetEnabled":true,"language":"python","sourceType":"notebook","isGpuEnabled":false}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"!pip install monai\n!python -c \"import monai\" || pip install -q \"monai-weekly[nibabel]\"\n!python -c \"import matplotlib\" || pip install -q matplotlib\n%matplotlib inline\n!pip install einops","metadata":{"execution":{"iopub.status.busy":"2024-02-20T18:39:39.97258Z","iopub.execute_input":"2024-02-20T18:39:39.972863Z","iopub.status.idle":"2024-02-20T18:40:31.864591Z","shell.execute_reply.started":"2024-02-20T18:39:39.972816Z","shell.execute_reply":"2024-02-20T18:40:31.863878Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# import torch.nn as nn\n# import itertools\n# from collections.abc import Sequence\n# from monai.networks.layers import Conv, trunc_normal_\n# class WindowAttention(nn.Module):\n\n# def __init__(\n# self,\n# dim: int,\n# num_heads: int,\n# window_size: Sequence[int]=(7,7,7),\n# qkv_bias: bool = False,\n# attn_drop: float = 0.0,\n# proj_drop: float = 0.0,\n# ) -> None:\n \n# super().__init__()\n# self.dim = dim\n# self.window_size = window_size\n# self.num_heads = num_heads\n# head_dim = dim // num_heads\n# self.scale = head_dim**-0.5\n# mesh_args = torch.meshgrid.__kwdefaults__\n\n# if len(self.window_size) == 3:\n# self.relative_position_bias_table = nn.Parameter(\n# torch.zeros(\n# (2 * self.window_size[0] - 1) * (2 * self.window_size[1] - 1) * (2 * self.window_size[2] - 1),\n# num_heads,\n# )\n# )\n# coords_d = torch.arange(self.window_size[0])\n# coords_h = torch.arange(self.window_size[1])\n# coords_w = torch.arange(self.window_size[2])\n# if mesh_args is not None:\n# coords = torch.stack(torch.meshgrid(coords_d, coords_h, coords_w, indexing=\"ij\"))\n# else:\n# coords = torch.stack(torch.meshgrid(coords_d, coords_h, coords_w))\n# coords_flatten = torch.flatten(coords, 1)\n# relative_coords = coords_flatten[:, :, None] - coords_flatten[:, None, :]\n# relative_coords = relative_coords.permute(1, 2, 0).contiguous()\n# relative_coords[:, :, 0] += self.window_size[0] - 1\n# relative_coords[:, :, 1] += self.window_size[1] - 1\n# relative_coords[:, :, 2] += self.window_size[2] - 1\n# relative_coords[:, :, 0] *= (2 * self.window_size[1] - 1) * (2 * self.window_size[2] - 1)\n# relative_coords[:, :, 1] *= 2 * self.window_size[2] - 1\n# elif len(self.window_size) == 2:\n# self.relative_position_bias_table = nn.Parameter(\n# torch.zeros((2 * window_size[0] - 1) * (2 * window_size[1] - 1), num_heads)\n# )\n# coords_h = torch.arange(self.window_size[0])\n# coords_w = torch.arange(self.window_size[1])\n# if mesh_args is not None:\n# coords = torch.stack(torch.meshgrid(coords_h, coords_w, indexing=\"ij\"))\n# else:\n# coords = torch.stack(torch.meshgrid(coords_h, coords_w))\n# coords_flatten = torch.flatten(coords, 1)\n# relative_coords = coords_flatten[:, :, None] - coords_flatten[:, None, :]\n# relative_coords = relative_coords.permute(1, 2, 0).contiguous()\n# relative_coords[:, :, 0] += self.window_size[0] - 1\n# relative_coords[:, :, 1] += self.window_size[1] - 1\n# relative_coords[:, :, 0] *= 2 * self.window_size[1] - 1\n\n# relative_position_index = relative_coords.sum(-1)\n# self.register_buffer(\"relative_position_index\", relative_position_index)\n# self.qkv = nn.Linear(dim, dim * 3, bias=qkv_bias)\n# self.attn_drop = nn.Dropout(attn_drop)\n# self.proj = nn.Linear(dim, dim)\n# self.proj_drop = nn.Dropout(proj_drop)\n# trunc_normal_(self.relative_position_bias_table, std=0.02)\n# self.softmax = nn.Softmax(dim=-1)\n\n# def forward(self, x, mask=None):\n# b, n, c = x.shape\n# print(x.shape, b, n, c, \"...........................................\")\n# qkv = self.qkv(x).reshape(b, n, 3, self.num_heads, c // self.num_heads).permute(2, 0, 3, 1, 4)\n# q, k, v = qkv[0], qkv[1], qkv[2]\n# print(q.shape, k.shape, v.shape, \"q, k, v\")\n# q = q * self.scale\n# attn = q @ k.transpose(-2, -1)\n# print(attn.shape, \"attn\")\n# relative_position_bias = self.relative_position_bias_table[\n# self.relative_position_index.clone()[:n, :n].reshape(-1)\n# ].reshape(n, n, -1)\n# relative_position_bias = relative_position_bias.permute(2, 0, 1).contiguous()\n# attn = attn + relative_position_bias.unsqueeze(0)\n# if mask is not None:\n# nw = mask.shape[0]\n# attn = attn.view(b // nw, nw, self.num_heads, n, n) + mask.unsqueeze(1).unsqueeze(0)\n# attn = attn.view(-1, self.num_heads, n, n)\n# attn = self.softmax(attn)\n# else:\n# attn = self.softmax(attn)\n\n# attn = self.attn_drop(attn).to(v.dtype)\n# print(attn.shape, \"attn.shape#########################################################\")\n# x = (attn @ v).transpose(1, 2).reshape(b, n, c)\n# print(x.shape, \"@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\")\n# x = self.proj(x)\n# print(x.shape, \"After proj\")\n# x = self.proj_drop(x)\n# return x\n","metadata":{"execution":{"iopub.status.busy":"2024-02-20T17:04:29.999537Z","iopub.execute_input":"2024-02-20T17:04:30.000462Z","iopub.status.idle":"2024-02-20T17:04:30.043348Z","shell.execute_reply.started":"2024-02-20T17:04:30.000411Z","shell.execute_reply":"2024-02-20T17:04:30.041357Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# import torch.nn as nn\n# import itertools\n# from collections.abc import Sequence\n# from monai.networks.layers import Conv, trunc_normal_\n# class WindowAttention(nn.Module):\n\n# def __init__(\n# self,\n# dim: int,\n# num_heads: int,\n# window_size: Sequence[int]=(7,7,7),\n# qkv_bias: bool = False,\n# attn_drop: float = 0.0,\n# proj_drop: float = 0.0,\n# ) -> None:\n \n# super().__init__()\n# self.dim = dim\n# self.window_size = window_size\n# self.num_heads = num_heads\n# head_dim = dim // num_heads\n# self.scale = head_dim**-0.5\n# mesh_args = torch.meshgrid.__kwdefaults__\n\n# if len(self.window_size) == 3:\n# self.relative_position_bias_table = nn.Parameter(\n# torch.zeros(\n# (2 * self.window_size[0] - 1) * (2 * self.window_size[1] - 1) * (2 * self.window_size[2] - 1),\n# num_heads,\n# )\n# )\n# coords_d = torch.arange(self.window_size[0])\n# coords_h = torch.arange(self.window_size[1])\n# coords_w = torch.arange(self.window_size[2])\n# if mesh_args is not None:\n# coords = torch.stack(torch.meshgrid(coords_d, coords_h, coords_w, indexing=\"ij\"))\n# else:\n# coords = torch.stack(torch.meshgrid(coords_d, coords_h, coords_w))\n# coords_flatten = torch.flatten(coords, 1)\n# relative_coords = coords_flatten[:, :, None] - coords_flatten[:, None, :]\n# relative_coords = relative_coords.permute(1, 2, 0).contiguous()\n# relative_coords[:, :, 0] += self.window_size[0] - 1\n# relative_coords[:, :, 1] += self.window_size[1] - 1\n# relative_coords[:, :, 2] += self.window_size[2] - 1\n# relative_coords[:, :, 0] *= (2 * self.window_size[1] - 1) * (2 * self.window_size[2] - 1)\n# relative_coords[:, :, 1] *= 2 * self.window_size[2] - 1\n# elif len(self.window_size) == 2:\n# self.relative_position_bias_table = nn.Parameter(\n# torch.zeros((2 * window_size[0] - 1) * (2 * window_size[1] - 1), num_heads)\n# )\n# coords_h = torch.arange(self.window_size[0])\n# coords_w = torch.arange(self.window_size[1])\n# if mesh_args is not None:\n# coords = torch.stack(torch.meshgrid(coords_h, coords_w, indexing=\"ij\"))\n# else:\n# coords = torch.stack(torch.meshgrid(coords_h, coords_w))\n# coords_flatten = torch.flatten(coords, 1)\n# relative_coords = coords_flatten[:, :, None] - coords_flatten[:, None, :]\n# relative_coords = relative_coords.permute(1, 2, 0).contiguous()\n# relative_coords[:, :, 0] += self.window_size[0] - 1\n# relative_coords[:, :, 1] += self.window_size[1] - 1\n# relative_coords[:, :, 0] *= 2 * self.window_size[1] - 1\n\n# relative_position_index = relative_coords.sum(-1)\n# self.register_buffer(\"relative_position_index\", relative_position_index)\n# self.qkv = nn.Linear(dim, dim * 3, bias=qkv_bias)\n# self.attn_drop = nn.Dropout(attn_drop)\n# self.proj = nn.Linear(dim, dim)\n# self.proj_drop = nn.Dropout(proj_drop)\n# trunc_normal_(self.relative_position_bias_table, std=0.02)\n# self.softmax = nn.Softmax(dim=-1)\n\n# def forward(self, x, mask=None):\n# print(x.shape , \"for enteting into window attention\")\n# b, n, c = x.shape\n# print(x.shape, b, n, c, \"...........................................\")\n# qkv = self.qkv(x).reshape(b, n, 3, self.num_heads, c // self.num_heads).permute(2, 0, 3, 1, 4)\n# q, k, v = qkv[0], qkv[1], qkv[2]\n# print(q.shape, k.shape, v.shape, \"q, k, v\")\n# q = q * self.scale\n# attn = q @ k.transpose(-2, -1)\n# print(attn.shape, \"attn\")\n# relative_position_bias = self.relative_position_bias_table[\n# self.relative_position_index.clone()[:n, :n].reshape(-1)\n# ].reshape(n, n, -1)\n# relative_position_bias = relative_position_bias.permute(2, 0, 1).contiguous()\n# attn = attn + relative_position_bias.unsqueeze(0)\n# if mask is not None:\n# nw = mask.shape[0]\n# attn = attn.view(b // nw, nw, self.num_heads, n, n) + mask.unsqueeze(1).unsqueeze(0)\n# attn = attn.view(-1, self.num_heads, n, n)\n# attn = self.softmax(attn)\n# else:\n# attn = self.softmax(attn)\n\n# attn = self.attn_drop(attn).to(v.dtype)\n# print(attn.shape, \"attn.shape#########################################################\")\n# x = (attn @ v).transpose(1, 2).reshape(b, n, c)\n# print(x.shape, \"@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\")\n# x = self.proj(x)\n# print(x.shape, \"After proj\")\n# x = self.proj_drop(x)\n# print(\"Exiting winfdoe attention\", x.shape)\n# return x\n","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","execution":{"iopub.status.busy":"2024-02-20T16:52:42.369455Z","iopub.execute_input":"2024-02-20T16:52:42.369944Z","iopub.status.idle":"2024-02-20T16:53:24.570042Z","shell.execute_reply.started":"2024-02-20T16:52:42.369898Z","shell.execute_reply":"2024-02-20T16:53:24.568732Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# import torch.nn as nn\n# import torch.nn.functional as F\n# from monai.networks.layers import Conv, trunc_normal_\n\n# class WindowAttention(nn.Module):\n# def __init__(self, dim, num_heads, window_size=(7,7,7), qkv_bias=False, attn_drop=0.0, proj_drop=0.0):\n# super().__init__()\n# self.dim = dim\n# self.window_size = window_size\n# self.num_heads = num_heads\n# head_dim = dim // num_heads\n# self.scale = head_dim**-0.5\n\n# # VSR module\n# self.avgpool = nn.AvgPool2d(window_size, stride=1)\n# self.leakyrelu = nn.LeakyReLU(negative_slope=0.01)\n# self.conv = ConvConv.CONV, 2\n\n# # Rest of the code remains the same\n\n# def forward(self, x, mask=None):\n# B, N, C = x.shape\n# x = x.view(B, N, self.num_heads, C // self.num_heads).permute(0, 2, 1, 3)\n\n# # VSR module\n# x = self.avgpool(x)\n# x = self.leakyrelu(x)\n# S_w, O_w = self.conv(x).chunk(2, dim=-1)\n# S_w = S_w.sigmoid()\n# O_w = O_w.tanh()\n\n# # Apply S_w and O_w to x here according to your methodology\n\n# # Rest of the code\n# relative_position_bias = self.relative_position_bias_table[self.relative_position_index.view(-1)].view(\n# b, n, n, -1\n# )\n# relative_position_bias = relative_position_bias.permute(3, 0, 1, 2).contiguous()\n\n# attn = attn + relative_position_bias.unsqueeze(0)\n# if mask is not None:\n# nw = mask.shape[0]\n# attn = attn.view(b // nw, nw, self.num_heads, n, n) + mask.unsqueeze(1).unsqueeze(0)\n# attn = attn.view(-1, self.num_heads, n, n)\n# attn = self.softmax(attn)\n# else:\n# attn = self.softmax(attn)\n\n# attn = self.attn_drop(attn).to(v.dtype)\n# x = (attn @ v).transpose(1, 2).reshape(b, n, c)\n# x = self.proj(x)\n# x = self.proj_drop(x)\n# return x\n","metadata":{"execution":{"iopub.status.busy":"2024-02-20T13:44:47.602161Z","iopub.execute_input":"2024-02-20T13:44:47.603309Z","iopub.status.idle":"2024-02-20T13:44:47.626018Z","shell.execute_reply.started":"2024-02-20T13:44:47.603259Z","shell.execute_reply":"2024-02-20T13:44:47.624496Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import torch\nimport torch.nn as nn\nimport torch.nn.functional as F\nfrom collections.abc import Sequence\n\nclass WindowAttention(nn.Module):\n def __init__(\n self,\n dim: int,\n num_heads: int,\n window_size: Sequence[int]=(7, 7, 7),\n qkv_bias: bool = False,\n attn_drop: float = 0.0,\n proj_drop: float = 0.0,\n ) -> None:\n super().__init__()\n self.dim = dim\n self.window_size = window_size\n self.num_heads = num_heads\n head_dim = dim // num_heads\n self.scale = head_dim ** -0.5\n\n # Define relative position bias table\n if len(self.window_size) == 3:\n self.relative_position_bias_table = nn.Parameter(\n torch.zeros(\n (2 * self.window_size[0] - 1) * (2 * self.window_size[1] - 1) * (2 * self.window_size[2] - 1),\n num_heads,\n )\n )\n coords_d = torch.arange(self.window_size[0])\n coords_h = torch.arange(self.window_size[1])\n coords_w = torch.arange(self.window_size[2])\n coords = torch.stack(torch.meshgrid(coords_d, coords_h, coords_w, indexing=\"ij\"))\n coords_flatten = torch.flatten(coords, 1)\n relative_coords = coords_flatten[:, :, None] - coords_flatten[:, None, :]\n relative_coords = relative_coords.permute(1, 2, 0).contiguous()\n relative_coords[:, :, 0] += self.window_size[0] - 1\n relative_coords[:, :, 1] += self.window_size[1] - 1\n relative_coords[:, :, 2] += self.window_size[2] - 1\n relative_coords[:, :, 0] *= (2 * self.window_size[1] - 1) * (2 * self.window_size[2] - 1)\n relative_coords[:, :, 1] *= 2 * self.window_size[2] - 1\n elif len(self.window_size) == 2:\n self.relative_position_bias_table = nn.Parameter(\n torch.zeros((2 * self.window_size[0] - 1) * (2 * self.window_size[1] - 1), num_heads)\n )\n coords_h = torch.arange(self.window_size[0])\n coords_w = torch.arange(self.window_size[1])\n coords = torch.stack(torch.meshgrid(coords_h, coords_w, indexing=\"ij\"))\n coords_flatten = torch.flatten(coords, 1)\n relative_coords = coords_flatten[:, :, None] - coords_flatten[:, None, :]\n relative_coords = relative_coords.permute(1, 2, 0).contiguous()\n relative_coords[:, :, 0] += self.window_size[0] - 1\n relative_coords[:, :, 1] += self.window_size[1] - 1\n relative_coords[:, :, 0] *= 2 * self.window_size[1] - 1\n\n relative_position_index = relative_coords.sum(-1)\n self.register_buffer(\"relative_position_index\", relative_position_index)\n\n # Define query, key, value linear layers\n self.qkv = nn.Linear(dim, dim * 3, bias=qkv_bias)\n self.attn_drop = nn.Dropout(attn_drop)\n self.proj = nn.Linear(dim, dim)\n self.proj_drop = nn.Dropout(proj_drop)\n self.softmax = nn.Softmax(dim=-1)\n\n def forward(self, x, mask=None):\n b, n, c = x.shape\n qkv = self.qkv(x).reshape(b, n, 3, self.num_heads, c // self.num_heads).permute(2, 0, 3, 1, 4)\n q, k, v = qkv[0], qkv[1], qkv[2]\n q = q * self.scale\n attn = q @ k.transpose(-2, -1)\n relative_position_bias = self.relative_position_bias_table[\n self.relative_position_index.clone()[:n, :n].reshape(-1)\n ].reshape(n, n, -1)\n relative_position_bias = relative_position_bias.permute(2, 0, 1).contiguous()\n attn = attn + relative_position_bias.unsqueeze(0)\n if mask is not None:\n nw = mask.shape[0]\n attn = attn.view(b // nw, nw, self.num_heads, n, n) + mask.unsqueeze(1).unsqueeze(0)\n attn = attn.view(-1, self.num_heads, n, n)\n attn = self.softmax(attn)\n else:\n attn = self.softmax(attn)\n\n attn = self.attn_drop(attn).to(v.dtype)\n x = (attn @ v).transpose(1, 2).reshape(b, n, c)\n x = self.proj(x)\n x = self.proj_drop(x)\n return x\n","metadata":{"execution":{"iopub.status.busy":"2024-02-20T18:40:34.098384Z","iopub.execute_input":"2024-02-20T18:40:34.098851Z","iopub.status.idle":"2024-02-20T18:40:35.183335Z","shell.execute_reply.started":"2024-02-20T18:40:34.098804Z","shell.execute_reply":"2024-02-20T18:40:35.182505Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# import torch\n# import torch.nn as nn\n\n# class VariedSizeWindowAttention(nn.Module):\n# def __init__(\n# self,\n# dim,\n# num_heads,\n# window_size,\n# qkv_bias=False,\n# attn_drop=0.0,\n# proj_drop=0.0,\n# ):\n# super().__init__()\n# self.dim = dim\n# self.num_heads = num_heads\n# self.window_size = window_size\n# self.qkv = nn.Linear(dim, dim * 3, bias=qkv_bias)\n# self.attn_drop = nn.Dropout(attn_drop)\n# self.proj = nn.Linear(dim, dim)\n# self.proj_drop = nn.Dropout(proj_drop)\n# self.softmax = nn.Softmax(dim=-1)\n\n# # Window regression module\n# self.window_regressor = nn.Linear(dim, 4) # Predicts size and location of the target window\n\n# def forward(self, x):\n# b, n, c = x.shape\n \n# # Perform linear transformation for query, key, and value\n# qkv = (\n# self.qkv(x)\n# .reshape(b, n, 3, self.num_heads, c // self.num_heads)\n# .permute(2, 0, 3, 1, 4)\n# )\n# q, k, v = qkv[0], qkv[1], qkv[2]\n# q = q * (self.dim // self.num_heads) ** -0.5\n\n# # Predict size and location of the target window\n# window_params = self.window_regressor(x) # Shape: (batch_size, sequence_length, 4)\n# target_window = self.get_target_window(window_params) # Get target window coordinates\n\n# # Sample key and value tokens from the target window\n# k, v = self.sample_from_window(k, v, target_window)\n\n# # Perform attention computation\n# attn = (q @ k.transpose(-2, -1)) * (self.window_size[0] ** -0.5) # Scale by window size\n# attn = attn.softmax(dim=-1)\n# attn = self.attn_drop(attn)\n\n# x = (attn @ v).transpose(1, 2).reshape(b, n, c)\n# x = self.proj(x)\n# x = self.proj_drop(x)\n# return x\n\n# def get_target_window(self, window_params):\n# # Extract predicted parameters for the target window\n# target_center_x = window_params[:, :, 0] # Predicted x-coordinate of the window center\n# target_center_y = window_params[:, :, 1] # Predicted y-coordinate of the window center\n# target_size_x = window_params[:, :, 2] # Predicted width of the window\n# target_size_y = window_params[:, :, 3] # Predicted height of the window\n\n# # Calculate coordinates of the target window\n# target_left = target_center_x - target_size_x / 2\n# target_right = target_center_x + target_size_x / 2\n# target_top = target_center_y - target_size_y / 2\n# target_bottom = target_center_y + target_size_y / 2\n\n# return target_left, target_right, target_top, target_bottom\n\n# def sample_from_window(self, k, v, target_window):\n# # Extract target window coordinates\n# target_left, target_right, target_top, target_bottom = target_window\n\n# # Sample key and value tokens from the target window\n# sampled_k = self.sample_tokens(k, target_left, target_right, target_top, target_bottom)\n# sampled_v = self.sample_tokens(v, target_left, target_right, target_top, target_bottom)\n\n# return sampled_k, sampled_v\n\n# def sample_tokens(self, tokens, left, right, top, bottom):\n# # Perform token sampling from the target window\n# batch_size, num_heads, seq_len, head_dim = tokens.size()\n# tokens = tokens.permute(0, 2, 1, 3).reshape(batch_size, seq_len, -1) # Reshape for sampling\n# sampled_tokens = []\n\n# for i in range(batch_size):\n# for j in range(seq_len):\n# token_x, token_y = j % self.window_size[0], j // self.window_size[0] # Token coordinates\n# if left[i, j] <= token_x <= right[i, j] and top[i, j] <= token_y <= bottom[i, j]:\n# sampled_tokens.append(tokens[i, j]) # Token is inside the target window\n\n# sampled_tokens = torch.stack(sampled_tokens, dim=0) if len(sampled_tokens) > 0 else torch.zeros(0) # Stack sampled tokens\n# return sampled_tokens.view(batch_size, num_heads, -1, head_dim).permute(0, 2, 1, 3) # Reshape back\n\n# import torch.nn as nn\n# # from models.TransBTS.IntmdSequential import IntermediateSequential \n# class IntermediateSequential(nn.Sequential):\n# def __init__(self, *args, return_intermediate=True):\n# super().__init__(*args)\n# self.return_intermediate = return_intermediate\n\n# def forward(self, input):\n# if not self.return_intermediate:\n# return super().forward(input)\n\n# intermediate_outputs = {}\n# output = input\n# for name, module in self.named_children():\n# output = intermediate_outputs[name] = module(output)\n\n# return output, intermediate_outputs\n# class SelfAttention(nn.Module):\n# def __init__(\n# self, dim, heads=8, qkv_bias=False, qk_scale=None, dropout_rate=0.0\n# ):\n# super().__init__()\n# self.num_heads = heads\n# head_dim = dim // heads\n# self.scale = qk_scale or head_dim ** -0.5\n\n# self.qkv = nn.Linear(dim, dim * 3, bias=qkv_bias)\n# self.attn_drop = nn.Dropout(dropout_rate)\n# self.proj = nn.Linear(dim, dim)\n# self.proj_drop = nn.Dropout(dropout_rate)\n\n# def forward(self, x):\n# B, N, C = x.shape\n# # print(B, N, C, \"B, N, C\")\n# qkv = (\n# self.qkv(x)\n# .reshape(B, N, 3, self.num_heads, C // self.num_heads)\n# .permute(2, 0, 3, 1, 4)\n# )\n# # print(x.shape)\n# # print(self.qkv(x).shape, \"self.qkv.shape\")\n \n# q, k, v = (\n# qkv[0],\n# qkv[1],\n# qkv[2],\n# ) # make torchscript happy (cannot use tensor as tuple)\n \n# # print(q.shape, k.shape, v.shape, \"q.shape, k.shape, v.shape\")\n# attn = (q @ k.transpose(-2, -1)) * self.scale\n# # print(attn.shape, \"attn.shape\")\n# attn = attn.softmax(dim=-1)\n# attn = self.attn_drop(attn)\n\n# x = (attn @ v).transpose(1, 2).reshape(B, N, C)\n# # print(x.shape, \"after multiplication with attn and V\")\n# x = self.proj(x)\n# # print(x.shape, \"after proj\")\n# x = self.proj_drop(x)\n# # print(x.shape, \"after proj drop\")\n# return x\n\n\n# class Residual(nn.Module):\n# def __init__(self, fn):\n# super().__init__()\n# self.fn = fn\n\n# def forward(self, x):\n# # print(\"In residual\", x.shape)\n# # print(self.fn, \"self.fn\")\n# return self.fn(x) + x\n\n\n# class PreNorm(nn.Module):\n# def __init__(self, dim, fn):\n# super().__init__()\n# self.norm = nn.LayerNorm(dim)\n# self.fn = fn\n\n# def forward(self, x):\n# # print(\"in PreNorm\", self.fn)\n# # print()\n# return self.fn(self.norm(x))\n\n\n# class PreNormDrop(nn.Module):\n# def __init__(self, dim, dropout_rate, fn):\n# super().__init__()\n# self.norm = nn.LayerNorm(dim)\n# self.dropout = nn.Dropout(p=dropout_rate)\n# self.fn = fn\n\n# def forward(self, x):\n# # print(\"In PreNormDrop\")\n# # print(self.fn, \"self.fn\")\n# # print(x.shape, \"x.shape\")\n# # print(self.norm(x).shape, \"self.norm(x)\")\n# return self.dropout(self.fn(self.norm(x)))\n\n\n# class FeedForward(nn.Module):\n# def __init__(self, dim, hidden_dim, dropout_rate):\n# super().__init__()\n# self.net = nn.Sequential(\n# nn.Linear(dim, hidden_dim),\n# nn.GELU(),\n# nn.Dropout(p=dropout_rate),\n# nn.Linear(hidden_dim, dim),\n# nn.Dropout(p=dropout_rate),\n# )\n\n# def forward(self, x):\n# # print(\"In feedforward\", x.shape)\n# # print(\"self.net(x)\", self.net(x).shape)\n# return self.net(x)\n\n# class TransformerModel(nn.Module):\n# def __init__(\n# self,\n# dim,\n# depth,\n# heads,\n# mlp_dim,\n# dropout_rate=0.1,\n# attn_dropout_rate=0.1,\n# window_size=(7, 7, 7)\n# ):\n# super().__init__()\n# self.dim = dim\n# self.depth = depth\n# self.heads = heads\n# self.mlp_dim = mlp_dim\n# self.dropout_rate = dropout_rate\n# self.attn_dropout_rate = attn_dropout_rate\n# self.window_size = window_size\n\n# self.layers = nn.ModuleList([\n# nn.ModuleList([\n# Residual(PreNormDrop(dim, dropout_rate,\n# VariedSizeWindowAttention(dim, num_heads=heads,\n# window_size=window_size,\n# attn_drop=attn_dropout_rate,\n# proj_drop=dropout_rate)),\n# ),\n# Residual(PreNorm(dim, FeedForward(dim, mlp_dim, dropout_rate))),\n# ])\n# for _ in range(depth)\n# ])\n\n# def forward(self, x):\n# for attn, ff in self.layers:\n# x = attn(x)\n# x = ff(x)\n# return x\n","metadata":{"execution":{"iopub.status.busy":"2024-02-20T13:31:34.802072Z","iopub.execute_input":"2024-02-20T13:31:34.802561Z","iopub.status.idle":"2024-02-20T13:31:34.819476Z","shell.execute_reply.started":"2024-02-20T13:31:34.802523Z","shell.execute_reply":"2024-02-20T13:31:34.818405Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# import torch.nn as nn\n# from collections.abc import Sequence\n# from monai.networks.layers import Conv, trunc_normal_\n\n# class WindowAttention(nn.Module):\n\n# def __init__(\n# self,\n# dim: int,\n# num_heads: int,\n# qkv_bias: bool = False,\n# attn_drop: float = 0.0,\n# proj_drop: float = 0.0,\n# ) -> None:\n \n# super().__init__()\n# self.dim = dim\n# self.num_heads = num_heads\n# head_dim = dim // num_heads\n# self.scale = head_dim**-0.5\n# self.qkv = nn.Linear(dim, dim * 3, bias=qkv_bias)\n# self.attn_drop = nn.Dropout(attn_drop)\n# self.proj = nn.Linear(dim, dim)\n# self.proj_drop = nn.Dropout(proj_drop)\n\n# def forward(self, x, mask=None):\n# b, n, c = x.shape\n# qkv = self.qkv(x).reshape(b, n, 3, self.num_heads, c // self.num_heads).permute(2, 0, 3, 1, 4)\n# q, k, v = qkv[0], qkv[1], qkv[2]\n# q = q * self.scale\n# attn = q @ k.transpose(-2, -1)\n\n# attn = self.softmax(attn)\n# attn = self.attn_drop(attn).to(v.dtype)\n\n# x = (attn @ v).transpose(1, 2).reshape(b, n, c)\n# x = self.proj(x)\n# x = self.proj_drop(x)\n# return x\n","metadata":{"execution":{"iopub.status.busy":"2024-02-20T13:31:35.142532Z","iopub.execute_input":"2024-02-20T13:31:35.143004Z","iopub.status.idle":"2024-02-20T13:31:35.152137Z","shell.execute_reply.started":"2024-02-20T13:31:35.142965Z","shell.execute_reply":"2024-02-20T13:31:35.150589Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# import torch\n# import torch.nn as nn\n# from monai.networks.layers import Conv, trunc_normal_\n# import torch.nn as nn\n# import torch.nn.functional as F\n# from collections.abc import Sequence\n# from monai.networks.layers import Conv, trunc_normal_\n\n# class VariedSizeWindowAttention(nn.Module):\n\n# def __init__(\n# self,\n# dim: int,\n# num_heads: int,\n# window_size: Sequence[int] = (7, 7), # Adjusted for 2D input\n# qkv_bias: bool = False,\n# attn_drop: float = 0.0,\n# proj_drop: float = 0.0,\n# ) -> None:\n# super().__init__()\n# self.dim = dim\n# self.num_heads = num_heads\n# self.window_size = window_size\n# head_dim = dim // num_heads\n# self.scale = head_dim**-0.5\n\n# # VSA components\n# self.vsr_module = nn.Sequential(\n# nn.AdaptiveAvgPool2d(output_size=window_size), # Adjusted for adaptive pooling\n# nn.LeakyReLU(),\n# nn.Conv2d(dim, 2 * num_heads, kernel_size=1, stride=1) # Predict scales and offsets\n# )\n# self.cpe = Conv(dim, dim, kernel_size=window_size, stride=1, padding=window_size[0] // 2)\n\n# # Window attention components\n# self.qkv = nn.Linear(dim, dim * 3, bias=qkv_bias)\n# self.attn_drop = nn.Dropout(attn_drop)\n# self.proj = nn.Linear(dim, dim)\n# self.proj_drop = nn.Dropout(proj_drop)\n\n# def forward(self, x, mask=None):\n# b, n, c = x.shape\n\n# # Varied-size window regression\n# sw, ow = self.vsr_module(x).chunk(2, dim=1)\n# sw = sw.sigmoid() # Scale to [0, 1]\n# ow = ow.tanh() # Offset to [-1, 1]\n\n# # Get key/value tokens and apply conditional position embedding\n# k, v = self.qkv(x).reshape(b, n, 3, self.num_heads, c // self.num_heads).permute(2, 0, 3, 1, 4)\n# k, v = self.cpe(k), self.cpe(v)\n\n# # Sample from varied-size windows\n# k_sampled = F.grid_sample(k, grid=(sw + ow).permute(0, 2, 3, 1), mode='bilinear', padding_mode='zeros') # Adaptive sampling\n# v_sampled = F.grid_sample(v, grid=(sw + ow).permute(0, 2, 3, 1), mode='bilinear', padding_mode='zeros')\n\n# # Attention calculation with sampled tokens\n# q = q * self.scale\n# attn = q @ k_sampled.transpose(-2, -1)\n# # ... (rest of the attention calculation and output)\n\n# attn = attn.softmax(dim=-1)\n# attn = self.attn_drop(attn)\n\n# x = (attn @ kv).transpose(1, 2).reshape(b, h * w, c)\n# x = x.permute(0, 2, 1).reshape(b, c, h, w) # Reshape back\n\n# # Projection\n# x = self.proj(x)\n# x = self.proj_drop(x)\n\n# return x\n","metadata":{"execution":{"iopub.status.busy":"2024-02-20T14:24:22.206651Z","iopub.execute_input":"2024-02-20T14:24:22.207186Z","iopub.status.idle":"2024-02-20T14:24:22.231444Z","shell.execute_reply.started":"2024-02-20T14:24:22.20715Z","shell.execute_reply":"2024-02-20T14:24:22.230408Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# import torch.nn as nn\n# import torch\n\n# class VariedSizeWindowRegression(nn.Module):\n# def __init__(self, in_channels, out_channels):\n# super().__init__()\n \n# self.avg_pool = nn.AvgPool2d(kernel_size=window_size, stride=window_size),\n# self.conv1x1 = nn.Conv2d(512, 4 * 2, kernel_size=1, stride=1)\n# self.leaky_relu = nn.LeakyReLU()\n\n# def forward(self, x):\n# x = self.avg_pool(x)\n# x = self.leaky_relu(x)\n# x = self.conv1x1(x)\n# return x\n\n# class WindowAttention(nn.Module):\n# def __init__(\n# self,\n# dim: int,\n# num_heads: int,\n# window_size: torch.Size,\n# qkv_bias: bool = False,\n# attn_drop: float = 0.0,\n# proj_drop: float = 0.0,\n# ) -> None:\n# super().__init__()\n# self.dim = dim\n# self.window_size = window_size\n# self.num_heads = num_heads\n# head_dim = dim // num_heads\n# self.scale = head_dim**-0.5\n# self.vs_regression = VariedSizeWindowRegression(dim, dim)\n# self.mlp = nn.Sequential(\n# nn.Linear(dim, dim),\n# nn.LeakyReLU(),\n# nn.Linear(dim, dim)\n# )\n# self.attn_drop = nn.Dropout(attn_drop)\n# self.proj = nn.Linear(dim, dim)\n# self.proj_drop = nn.Dropout(proj_drop)\n# self.softmax = nn.Softmax(dim=-1)\n\n# def forward(self, x, mask=None):\n# print(x.shape, \"for entering into window attention\")\n# b, n, c = x.shape\n# print(x.shape, b, n, c, \"...........................................\")\n \n# # Preprocessing with Varied-size Window Regression\n# x = self.vs_regression(x)\n \n# qkv = self.mlp(x).reshape(b, n, 3, self.num_heads, c // self.num_heads).permute(2, 0, 3, 1, 4)\n# q, k, v = qkv[0], qkv[1], qkv[2]\n# print(q.shape, k.shape, v.shape, \"q, k, v\")\n# q = q * self.scale\n# attn = q @ k.transpose(-2, -1)\n# print(attn.shape, \"attn\")\n \n# if mask is not None:\n# nw = mask.shape[0]\n# attn = attn.view(b // nw, nw, self.num_heads, n, n) + mask.unsqueeze(1).unsqueeze(0)\n# attn = attn.view(-1, self.num_heads, n, n)\n# attn = self.softmax(attn)\n# else:\n# attn = self.softmax(attn)\n\n# attn = self.attn_drop(attn).to(v.dtype)\n# print(attn.shape, \"attn.shape#########################################################\")\n# x = (attn @ v).transpose(1, 2).reshape(b, n, c)\n# print(x.shape, \"@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\")\n# x = self.proj(x)\n# print(x.shape, \"After proj\")\n# x = self.proj_drop(x)\n# print(\"Exiting window attention\", x.shape)\n# return x\n\n# # Example usage:\n# input_shape = (32, 8, 512)\n# num_heads = 4\n# window_size = (7, 7, 7)\n# attention = WindowAttention(dim=512, num_heads=num_heads, window_size=window_size)\n# input_tensor = torch.randn(input_shape)\n# output = attention(input_tensor)\n","metadata":{"execution":{"iopub.status.busy":"2024-02-20T16:54:37.344417Z","iopub.execute_input":"2024-02-20T16:54:37.344805Z","iopub.status.idle":"2024-02-20T16:54:37.842961Z","shell.execute_reply.started":"2024-02-20T16:54:37.344774Z","shell.execute_reply":"2024-02-20T16:54:37.841311Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# import torch\n# import torch.nn as nn\n# import torch.nn.functional as F\n# from collections.abc import Sequence\n# from monai.networks.layers import Conv, trunc_normal_\n\n# class WindowAttention(nn.Module):\n# def __init__(\n# self,\n# dim: int,\n# num_heads: int,\n# window_size: Sequence[int] = (7, 7), # Change default window_size to 2D\n# qkv_bias: bool = False,\n# attn_drop: float = 0.0,\n# proj_drop: float = 0.0,\n# ) -> None:\n# super().__init__()\n# self.dim = dim\n# self.window_size = window_size\n# self.num_heads = num_heads\n# head_dim = dim // num_heads\n# self.scale = head_dim ** -0.5\n# mesh_args = torch.meshgrid.__kwdefaults__\n\n# # Define sampling coefficients layers\n# self.sampling_coefficients = nn.Sequential(\n# nn.AvgPool2d(kernel_size=(window_size[0], window_size[1])), # Omit stride for no striding\n# nn.LeakyReLU(),\n# nn.Conv2d(36, self.num_heads * 2, kernel_size=1, stride=1)\n# )\n \n# if len(self.window_size) == 3:\n# coords_d = torch.arange(self.window_size[0])\n# coords_h = torch.arange(self.window_size[1])\n# coords_w = torch.arange(self.window_size[2])\n# if mesh_args is not None:\n# coords = torch.stack(torch.meshgrid(coords_d, coords_h, coords_w, indexing=\"ij\"))\n# else:\n# coords = torch.stack(torch.meshgrid(coords_d, coords_h, coords_w))\n# coords_flatten = torch.flatten(coords, 1)\n# relative_coords = coords_flatten[:, :, None] - coords_flatten[:, None, :]\n# relative_coords = relative_coords.permute(1, 2, 0).contiguous()\n# relative_coords[:, :, 0] += self.window_size[0] - 1\n# relative_coords[:, :, 1] += self.window_size[1] - 1\n# relative_coords[:, :, 2] += self.window_size[2] - 1\n# relative_coords[:, :, 0] *= (2 * self.window_size[1] - 1) * (2 * self.window_size[2] - 1)\n# relative_coords[:, :, 1] *= 2 * self.window_size[2] - 1\n# elif len(self.window_size) == 2:\n# coords_h = torch.arange(self.window_size[0])\n# coords_w = torch.arange(self.window_size[1])\n# if mesh_args is not None:\n# coords = torch.stack(torch.meshgrid(coords_h, coords_w, indexing=\"ij\"))\n# else:\n# coords = torch.stack(torch.meshgrid(coords_h, coords_w))\n# coords_flatten = torch.flatten(coords, 1)\n# relative_coords = coords_flatten[:, :, None] - coords_flatten[:, None, :]\n# relative_coords = relative_coords.permute(1, 2, 0).contiguous()\n# relative_coords[:, :, 0] += self.window_size[0] - 1\n# relative_coords[:, :, 1] += self.window_size[1] - 1\n# relative_coords[:, :, 0] *= 2 * self.window_size[1] - 1\n\n# relative_position_index = relative_coords.sum(-1)\n# self.register_buffer(\"relative_position_index\", relative_position_index)\n# self.qkv = nn.Linear(dim, dim * 3, bias=qkv_bias)\n# self.attn_drop = nn.Dropout(attn_drop)\n# self.proj = nn.Linear(dim, dim)\n# self.proj_drop = nn.Dropout(proj_drop)\n# self.relative_position_bias_table = nn.Parameter(\n# torch.zeros_like(relative_position_index, dtype=torch.float32)\n# )\n# trunc_normal_(self.relative_position_bias_table, std=0.02)\n# self.softmax = nn.Softmax(dim=-1)\n\n# # Initialize parameters\n# self.apply(self._init_weights)\n\n# def forward(self, x, mask=None):\n# b, n, c = x.shape\n\n# # Compute sampling coefficients\n# sampling_coeffs = self.sampling_coefficients(x)\n\n# # Compute sampling offsets and scales if applicable\n# if hasattr(self, 'sampling_offsets'):\n# sampling_offsets = self.sampling_offsets(x)\n# sampling_scales = self.sampling_scales(x)\n# # Adjust coordinates using sampling offsets and scales\n# # ...\n\n# qkv = self.qkv(x).reshape(b, n, 3, self.num_heads, c // self.num_heads).permute(2, 0, 3, 1, 4)\n# q, k, v = qkv[0], qkv[1], qkv[2]\n# q = q * self.scale\n# attn = q @ k.transpose(-2, -1)\n# relative_position_bias = self.relative_position_bias_table[\n# self.relative_position_index.clone()[:n, :n].reshape(-1)\n# ].reshape(n, n, -1)\n# relative_position_bias = relative_position_bias.permute(2, 0, 1).contiguous()\n# attn = attn + relative_position_bias.unsqueeze(0)\n# if mask is not None:\n# nw = mask.shape[0]\n# attn = attn.view(b // nw, nw, self.num_heads, n, n) + mask.unsqueeze(1).unsqueeze(0)\n# attn = attn.view(-1, self.num_heads, n, n)\n# attn = self.softmax(attn)\n# else:\n# attn = self.softmax(attn)\n\n# attn = self.attn_drop(attn).to(v.dtype)\n# x = (attn @ v).transpose(1, 2).reshape(b, n, c)\n# x = self.proj(x)\n# x = self.proj_drop(x)\n# return x\n\n# def _compute_relative_position_index(self):\n# if len(self.window_size) == 3:\n# coords_d = torch.arange(self.window_size[0])\n# coords_h = torch.arange(self.window_size[1])\n# coords_w = torch.arange(self.window_size[2])\n# coords = torch.stack(torch.meshgrid(coords_d, coords_h, coords_w, indexing=\"ij\"))\n# coords_flatten = torch.flatten(coords, 1)\n# relative_coords = coords_flatten[:, :, None] - coords_flatten[:, None, :]\n# relative_coords = relative_coords.permute(1, 2, 0).contiguous()\n# relative_coords[:, :, 0] += self.window_size[0] - 1\n# relative_coords[:, :, 1] += self.window_size[1] - 1\n# relative_coords[:, :, 2] += self.window_size[2] - 1\n# relative_coords[:, :, 0] *= (2 * self.window_size[1] - 1) * (2 * self.window_size[2] - 1)\n# relative_coords[:, :, 1] *= 2 * self.window_size[2] - 1\n# elif len(self.window_size) == 2:\n# coords_h = torch.arange(self.window_size[0])\n# coords_w = torch.arange(self.window_size[1])\n# coords = torch.stack(torch.meshgrid(coords_h, coords_w, indexing=\"ij\"))\n# coords_flatten = torch.flatten(coords, 1)\n# relative_coords = coords_flatten[:, :, None] - coords_flatten[:, None, :]\n# relative_coords = relative_coords.permute(1, 2, 0).contiguous()\n# relative_coords[:, :, 0] += self.window_size[0] - 1\n# relative_coords[:, :, 1] += self.window_size[1] - 1\n# relative_coords[:, :, 0] *= 2 * self.window_size[1] - 1\n# relative_position_index = relative_coords.sum(-1)\n# return relative_position_index\n\n# def _init_weights(self, m):\n# if isinstance(m, nn.Linear):\n# trunc_normal_(m.weight, std=0.02)\n# if isinstance(m, nn.Linear) and m.bias is not None:\n# nn.init.constant_(m.bias, 0)\n# elif isinstance(m, nn.Conv2d):\n# trunc_normal_(m.weight, std=0.02)\n# if isinstance(m, nn.Conv2d) and m.bias is not None:\n# nn.init.constant_(m.bias, 0)\n# # Example usage:\n# input_shape = (32, 8, 512)\n# num_heads = 4\n# window_size = (7, 7, 7)\n# attention = WindowAttention(dim=512, num_heads=num_heads, window_size=window_size)\n# input_tensor = torch.randn(input_shape)\n# output = attention(input_tensor)\n","metadata":{"execution":{"iopub.status.busy":"2024-02-20T15:40:55.896429Z","iopub.execute_input":"2024-02-20T15:40:55.897091Z","iopub.status.idle":"2024-02-20T15:40:56.347935Z","shell.execute_reply.started":"2024-02-20T15:40:55.897039Z","shell.execute_reply":"2024-02-20T15:40:56.345824Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import torch.nn as nn\n# from models.TransBTS.IntmdSequential import IntermediateSequential \nclass IntermediateSequential(nn.Sequential):\n def __init__(self, *args, return_intermediate=True):\n super().__init__(*args)\n self.return_intermediate = return_intermediate\n\n def forward(self, input):\n if not self.return_intermediate:\n return super().forward(input)\n\n intermediate_outputs = {}\n output = input\n for name, module in self.named_children():\n output = intermediate_outputs[name] = module(output)\n\n return output, intermediate_outputs\nclass SelfAttention(nn.Module):\n def __init__(\n self, dim, heads=8, qkv_bias=False, qk_scale=None, dropout_rate=0.0\n ):\n super().__init__()\n self.num_heads = heads\n head_dim = dim // heads\n self.scale = qk_scale or head_dim ** -0.5\n\n self.qkv = nn.Linear(dim, dim * 3, bias=qkv_bias)\n self.attn_drop = nn.Dropout(dropout_rate)\n self.proj = nn.Linear(dim, dim)\n self.proj_drop = nn.Dropout(dropout_rate)\n\n def forward(self, x):\n B, N, C = x.shape\n# print(B, N, C, \"B, N, C\")\n qkv = (\n self.qkv(x)\n .reshape(B, N, 3, self.num_heads, C // self.num_heads)\n .permute(2, 0, 3, 1, 4)\n )\n# print(x.shape)\n# print(self.qkv(x).shape, \"self.qkv.shape\")\n \n q, k, v = (\n qkv[0],\n qkv[1],\n qkv[2],\n ) # make torchscript happy (cannot use tensor as tuple)\n \n# print(q.shape, k.shape, v.shape, \"q.shape, k.shape, v.shape\")\n attn = (q @ k.transpose(-2, -1)) * self.scale\n# print(attn.shape, \"attn.shape\")\n attn = attn.softmax(dim=-1)\n attn = self.attn_drop(attn)\n\n x = (attn @ v).transpose(1, 2).reshape(B, N, C)\n# print(x.shape, \"after multiplication with attn and V\")\n x = self.proj(x)\n# print(x.shape, \"after proj\")\n x = self.proj_drop(x)\n# print(x.shape, \"after proj drop\")\n return x\n\n\nclass Residual(nn.Module):\n def __init__(self, fn):\n super().__init__()\n self.fn = fn\n\n def forward(self, x):\n# print(\"In residual\", x.shape)\n# print(self.fn, \"self.fn\")\n return self.fn(x) + x\n\n\nclass PreNorm(nn.Module):\n def __init__(self, dim, fn):\n super().__init__()\n self.norm = nn.LayerNorm(dim)\n self.fn = fn\n\n def forward(self, x):\n# print(\"in PreNorm\", self.fn)\n# print()\n return self.fn(self.norm(x))\n\n\nclass PreNormDrop(nn.Module):\n def __init__(self, dim, dropout_rate, fn):\n super().__init__()\n self.norm = nn.LayerNorm(dim)\n self.dropout = nn.Dropout(p=dropout_rate)\n self.fn = fn\n\n def forward(self, x):\n# print(\"In PreNormDrop\")\n# print(self.fn, \"self.fn\")\n# print(x.shape, \"x.shape\")\n# print(self.norm(x).shape, \"self.norm(x)\")\n return self.dropout(self.fn(self.norm(x)))\n\n\nclass FeedForward(nn.Module):\n def __init__(self, dim, hidden_dim, dropout_rate):\n super().__init__()\n self.net = nn.Sequential(\n nn.Linear(dim, hidden_dim),\n nn.GELU(),\n nn.Dropout(p=dropout_rate),\n nn.Linear(hidden_dim, dim),\n nn.Dropout(p=dropout_rate),\n )\n\n def forward(self, x):\n# print(\"In feedforward\", x.shape)\n# print(\"self.net(x)\", self.net(x).shape)\n return self.net(x)\n\nclass TransformerModel(nn.Module):\n def __init__(\n self,\n dim,\n depth,\n heads,\n mlp_dim,\n dropout_rate=0.1,\n attn_dropout_rate=0.1,\n ):\n super().__init__()\n layers = []\n for i in range(depth):\n # Alternate between SelfAttention and WindowAttention\n if i % 2 == 0:\n attention_layer = SelfAttention(dim, heads=heads, dropout_rate=attn_dropout_rate)\n else:\n attention_layer = WindowAttention(\n dim=dim,\n num_heads=heads,\n window_size=(7, 7, 7),\n qkv_bias=True,\n attn_drop=attn_dropout_rate,\n proj_drop=0.1\n )\n\n layers.extend(\n [\n Residual(\n PreNormDrop(\n dim,\n dropout_rate,\n attention_layer\n )\n ),\n Residual(\n PreNorm(dim, FeedForward(dim, mlp_dim, dropout_rate))\n ),\n ]\n )\n self.net = IntermediateSequential(*layers)\n\n def forward(self, x):\n return self.net(x)\n\nmodel = TransformerModel(512, 4, 8, 4096, 0.1, 0.1)\nprint(model)\n\n","metadata":{"execution":{"iopub.status.busy":"2024-02-20T18:40:44.413749Z","iopub.execute_input":"2024-02-20T18:40:44.414614Z","iopub.status.idle":"2024-02-20T18:40:44.692396Z","shell.execute_reply.started":"2024-02-20T18:40:44.414582Z","shell.execute_reply":"2024-02-20T18:40:44.691582Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import torch\nimport torch.nn as nn\nfrom monai.utils import deprecated_arg, ensure_tuple_rep, optional_import\nfrom monai.utils.module import look_up_option\nimport numpy as np\nrearrange, _ = optional_import(\"einops\", name=\"rearrange\")\nRearrange, _ = optional_import(\"einops.layers.torch\", name=\"Rearrange\")\nimport torch\nimport torch.nn as nn\n\n\nclass ConvolutionalVisionTransformer(nn.Module):\n def __init__(self, in_channels, num_classes, patch_size=16, dim=8, num_layers=6, num_heads=8, dim_feedforward=2048, dropout=0.1):\n super(ConvolutionalVisionTransformer, self).__init__()\n\n # Patch embedding layer\n self.patch_embedding = nn.Conv3d(in_channels, dim, kernel_size=patch_size, stride=patch_size)\n self.dim=dim\n # Calculate number of patches\n self.num_patches = (int((128 - patch_size) / patch_size) + 1) ** 3\n\n # Positional embedding (corrected shape)\n self.positional_embedding = nn.Parameter(torch.zeros(1, self.num_patches, dim)) # Removed extra dimension\n\n\n # Convolutional layers\n self.conv_layers = nn.Sequential(\n nn.Conv3d(dim, dim, kernel_size=(1, 3, 3), stride=(1, 1, 1), padding=(0, 1, 1)),\n nn.BatchNorm3d(dim),\n nn.ReLU(),\n nn.Conv3d(dim, dim, kernel_size=(1, 3, 3), stride=(1, 1, 1), padding=(0, 1, 1)),\n nn.BatchNorm3d(dim),\n nn.ReLU(),\n nn.Conv3d(dim, dim, kernel_size=(1, 3, 3), stride=(1, 1, 1), padding=(0, 1, 1)),\n nn.BatchNorm3d(dim),\n nn.ReLU()\n )\n# self.transformer = nn.TransformerEncoder(\n# nn.TransformerEncoderLayer(d_model=dim, nhead=num_heads, dim_feedforward=dim_feedforward, dropout=dropout),\n# num_layers=num_layers\n# )\n self.transformer2=TransformerModel(512,4,8,4096,0.1,0.1)\n# num_layers=4,\n# num_heads=8,\n# hidden_dim=4096,\n# dropout_rate=0.1,\n# attn_dropout_rate=0.1)\n self.classification_head = nn.Linear(512, num_classes)\n\n def forward(self, x):\n x = self.patch_embedding(x) \n x = x.flatten(2).transpose(1, 2)\n x = x + self.positional_embedding[:, :x.size(1)]\n print(x.shape)\n print(\"shape of x before passing into conv layers\",x.shape)\n# x = torch.tensor([x[0][0], x[0][2], x[0][1]])\n# print(x.shape)\n x = x.permute(0, 2, 1)\n print(x.shape)\n# x = x.view(x.shape[0], x.shape[1], x.size(2), 1, 1) \n print(x.shape)\n \n# x = self.conv_layers(x)\n# x = x.flatten(2).transpose(1, 2)\n# x=x.squeeze(-1).squeeze(-1)\n print(\"before transformer\",x.shape)\n x_shape = x.size()\n print(x_shape, \"In basic layer \")\n# if len(x_shape) == 5:\n# b, c, d, h, w = x_shape\n# window_size, shift_size = get_window_size((d, h, w), (7,7,7), (3,3,3))\n# x = rearrange(x, \"b c d h w -> b d h w c\")\n# print(b, c, h, w, \"b, c, h, w values\", (7,7,7), (3,3,3), \"self.window_size, self.shift_size\")\n \n# dp = int(np.ceil(d / window_size[0])) * window_size[0]\n# hp = int(np.ceil(h / window_size[1])) * window_size[1]\n# wp = int(np.ceil(w / window_size[2])) * window_size[2]\n# print(window_size[0], window_si\n# ze[1], h, w, d, hp, wp, dp, \n# \"window_size[0], window_size[1], h, w, d, hp, wp, dp\")\n \n# attn_mask = compute_mask([dp, hp, wp], window_size, shift_size, x.device)\n \n# x = x.view(b, d, h, w, -1)\n# x = rearrange(x, \"b d h w c -> b c d h w\")\n x, _ = self.transformer2(x) # Unpack the tuple\n x = x.mean(dim=1) # Now you can take the mean\n logits = self.classification_head(x)\n return logits\n\n\nmodel = ConvolutionalVisionTransformer(1, 14)\n\n\ndevice = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\nmodel = model.to(device)\nprint(model)\n# Create sample input tensors\nbatch_images = torch.randn(32, 1, 128, 128,128).to(device)\n\n# # Forward pass\nclassification_outputs = model(batch_images)\nprint(classification_outputs.shape) # Output shape: (32, 14)\nprint(\"window attention \",classification_outputs)\n","metadata":{"execution":{"iopub.status.busy":"2024-02-20T18:40:50.68134Z","iopub.execute_input":"2024-02-20T18:40:50.68163Z","iopub.status.idle":"2024-02-20T18:41:09.593973Z","shell.execute_reply.started":"2024-02-20T18:40:50.681604Z","shell.execute_reply":"2024-02-20T18:41:09.593307Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# import torch.nn as nn\n# import torch\n# class VariedSizeWindowRegression(nn.Module):\n# def __init__(self, in_channels, out_channels, window_size=(7, 7, 7)):\n# super().__init__()\n# self.avg_pool = nn.AdaptiveAvgPool3d(1)\n# self.conv1x1 = nn.Conv3d(in_channels, out_channels, kernel_size=1, stride=1)\n# self.leaky_relu = nn.LeakyReLU()\n# self.window_size = window_size\n\n# def forward(self, x):\n# b, c, d, h, w = x.shape\n# # Reshape the input tensor to make it compatible with the Conv3d layer\n# x = x.reshape(b * c, 1, d, h, w)\n# x = self.conv1x1(x)\n# x = self.leaky_relu(x)\n# # Reshape back to the original shape\n# x = x.reshape(b, c, -1)\n# return x\n\n\n\n# class WindowAttention(nn.Module):\n# def __init__(\n# self,\n# dim: int,\n# num_heads: int,\n# window_size: torch.Size,\n# qkv_bias: bool = False,\n# attn_drop: float = 0.0,\n# proj_drop: float = 0.0,\n# ) -> None:\n# super().__init__()\n# self.dim = dim\n# self.window_size = window_size\n# self.num_heads = num_heads\n# head_dim = dim // num_heads\n# self.scale = head_dim**-0.5\n# self.vs_regression = VariedSizeWindowRegression(dim, dim)\n# self.mlp = nn.Sequential(\n# nn.Linear(dim, dim),\n# nn.LeakyReLU(),\n# nn.Linear(dim, dim)\n# )\n# self.attn_drop = nn.Dropout(attn_drop)\n# self.proj = nn.Linear(dim, dim)\n# self.proj_drop = nn.Dropout(proj_drop)\n# self.softmax = nn.Softmax(dim=-1)\n\n# def forward(self, x, mask=None):\n# print(x.shape, \"for entering into window attention\")\n# b, n, c = x.shape\n# print(x.shape, b, n, c, \"...........................................\")\n \n# # Preprocessing with Varied-size Window Regression\n# x = self.vs_regression(x)\n \n# qkv = self.mlp(x).reshape(b, n, 3, self.num_heads, c // self.num_heads).permute(2, 0, 3, 1, 4)\n# q, k, v = qkv[0], qkv[1], qkv[2]\n# print(q.shape, k.shape, v.shape, \"q, k, v\")\n# q = q * self.scale\n# attn = q @ k.transpose(-2, -1)\n# print(attn.shape, \"attn\")\n \n# if mask is not None:\n# nw = mask.shape[0]\n# attn = attn.view(b // nw, nw, self.num_heads, n, n) + mask.unsqueeze(1).unsqueeze(0)\n# attn = attn.view(-1, self.num_heads, n, n)\n# attn = self.softmax(attn)\n# else:\n# attn = self.softmax(attn)\n\n# attn = self.attn_drop(attn).to(v.dtype)\n# print(attn.shape, \"attn.shape#########################################################\")\n# x = (attn @ v).transpose(1, 2).reshape(b, n, c)\n# print(x.shape, \"@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@\")\n# x = self.proj(x)\n# print(x.shape, \"After proj\")\n# x = self.proj_drop(x)\n# print(\"Exiting window attention\", x.shape)\n# return x\n\n# # Example usage:\n# input_shape = (32, 8, 512)\n# num_heads = 4\n# window_size = (7, 7, 7)\n# attention = WindowAttention(dim=input_shape[-1], num_heads=num_heads, window_size=window_size)\n# input_tensor = torch.randn(input_shape)\n# output = attention(input_tensor)\n","metadata":{"execution":{"iopub.status.busy":"2024-02-20T14:21:05.722234Z","iopub.execute_input":"2024-02-20T14:21:05.722678Z","iopub.status.idle":"2024-02-20T14:21:06.072419Z","shell.execute_reply.started":"2024-02-20T14:21:05.722644Z","shell.execute_reply":"2024-02-20T14:21:06.069971Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# import os\n# import pandas as pd\n# import numpy as np\n# import nibabel as nib\n# import torch\n# import torch.nn as nn\n# import torch.optim as optim\n# from torch.utils.data import Dataset, DataLoader\n# from torchvision import transforms\n# import numpy as np\n# from sklearn import metrics\n# from sklearn.metrics import precision_recall_fscore_support\n# from sklearn.model_selection import train_test_split\n# from scipy.ndimage import zoom\n# from sklearn.metrics import confusion_matrix\n# from sklearn.metrics import accuracy_score, precision_score\n# # Create a function to move data to the device\n# def move_data_to_device(data, device):\n# return data.to(torch.float32).to(device)\n\n# class CustomDataset(Dataset):\n# def __init__(self, image_paths, labels, transform=None):\n# self.image_paths = image_paths\n# self.labels = labels\n# self.transform = transform\n\n# def __len__(self):\n# return len(self.image_paths)\n\n# def __getitem__(self, idx):\n# image_path = self.image_paths[idx]\n\n# # Load the 3D NIfTI image using nibabel\n# image = nib.load(image_path).get_fdata()\n\n# # Apply transformations if provided to the image\n# if self.transform:\n# image = self.transform(image)\n\n# label = torch.tensor(self.labels[idx], dtype=torch.float32)\n \n# return image, label\n\n# import torch.nn.functional as F\n\n# # Function to resize NIfTI data\n# def resize_nifti(nifti_data, target_shape):\n# factors = (target_shape[0] / nifti_data.shape[0],\n# target_shape[1] / nifti_data.shape[1],\n# target_shape[2] / nifti_data.shape[2])\n# resized_data = zoom(nifti_data, factors, order=3) # Cubic interpolation (higher quality)\n# return resized_data\n\n# # Paths and settings\n# csv_file = '/kaggle/input/unhealthy-csv-file/combined_data (6).csv' # Update with the correct path\n# batch_size = 32\n# num_workers = 4 # Number of CPU cores to use for data loading\n# num_classes = 14 # Number of classes\n# desired_shape = (128, 128, 128)\n# device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\n\n# # Define transformations if needed\n# transform = transforms.Compose([\n# transforms.ToTensor(), # Convert to tensor\n# # Add more transformations if necessary\n# ])\n\n# # Load the CSV file\n# data = pd.read_csv(csv_file).head(1280)\n\n# # Remove the extra space from the column name\n# data.columns = data.columns.str.strip()\n\n# # Assuming 'data' is your DataFrame\n# data_length = len(data)\n# print(\"Length of DataFrame:\", data_length)\n\n# # Split the data into training, validation, and test sets\n# train_data, temp_data = train_test_split(data, test_size=0.2, random_state=42)\n# val_data, test_data = train_test_split(temp_data, test_size=0.5, random_state=42)\n\n# # Set the display option to show all rows\n# pd.set_option('display.max_rows', None)\n\n# index_values = train_data.index.values\n\n# # Reset the display option to its default value (if needed)\n# pd.reset_option('display.max_rows')\n\n# # Extract file paths and labels from the data\n# train_paths = train_data['file_path'].values\n# train_labels = train_data[['bowel_healthy','bowel_injury','extravasation_healthy','extravasation_injury','kidney_healthy','kidney_low','kidney_high','liver_healthy','liver_low','liver_high','spleen_healthy','spleen_low','spleen_high','any_injury']].values\n# # print(\"train paths\", train_paths)\n\n\n# val_paths = val_data['file_path'].values\n# val_labels = val_data[['bowel_healthy','bowel_injury','extravasation_healthy','extravasation_injury','kidney_healthy','kidney_low','kidney_high','liver_healthy','liver_low','liver_high','spleen_healthy','spleen_low','spleen_high','any_injury']].values\n# # print(val_paths)\n\n\n# test_paths = test_data['file_path'].values\n# test_labels = test_data[['bowel_healthy','bowel_injury','extravasation_healthy','extravasation_injury','kidney_healthy','kidney_low','kidney_high','liver_healthy','liver_low','liver_high','spleen_healthy','spleen_low','spleen_high','any_injury']].values\n# # print(test_paths)\n\n\n# # Instantiate the datasets\n# train_dataset = CustomDataset(train_paths, train_labels, transform=transform)\n# print('len of train_dataset', len(train_dataset))\n# val_dataset = CustomDataset(val_paths, val_labels, transform=transform)\n# test_dataset = CustomDataset(test_paths, test_labels, transform=transform)\n\n# # Instantiate the data loaders\n# train_loader = DataLoader(train_dataset, batch_size=batch_size, shuffle=True, num_workers=num_workers, drop_last=True)\n# print('train_loader', len(train_loader))\n# val_loader = DataLoader(val_dataset, batch_size=batch_size, shuffle=False, num_workers=num_workers)\n# test_loader = DataLoader(test_dataset, batch_size=batch_size, shuffle=False, num_workers=num_workers)\n# print(train_loader)\n\n# # Instantiate the model with the appropriate number of classes for classification\n# in_channels = 1 # Input channels (e.g., for grayscale images or volumes)\n# num_classes_classification = 14 # Number of classes for classification\n# model_class = ConvolutionalVisionTransformer(in_channels, num_classes_classification)\n\n# # Count the number of parameters\n# total_params_class = sum(p.numel() for p in model_class.parameters())\n# print(f\"Total Trainable Parameters for Classification: {total_params_class}\")\n\n# # Define loss function and optimizer\n# class_criterion = nn.CrossEntropyLoss() # Binary Cross-Entropy loss for classification\n# class_optimizer = optim.Adam(model_class.parameters(), lr=0.001)\n\n# # Training loop\n# # Training loop\n# class_labels = ['bowel', 'extravasation', 'kidney', 'liver', 'spleen', 'any_injury']\n\n\n# # Training loop\n# num_epochs = 500\n# for epoch in range(num_epochs):\n# model_class.train()\n# running_loss = 0.0\n# correct_train = 0\n# total_train = 0\n# batch_number=0\n# all_predicted_labels = []\n# all_batch_labels = []\n# for batch_images, batch_labels in train_loader:\n# batch_number=batch_number+1\n# print(batch_number)\n# # Move data to the GPU if available\n# batch_images = batch_images.to(torch.float32).to(device)\n# batch_labels = batch_labels.to(torch.float32).to(device)\n# print(batch_labels.shape)\n# print(batch_labels)\n\n# # Assuming batch_images has shape (batch_size, num_frames, num_channels, height, width)\n# batch_images = batch_images.unsqueeze(1) # Add a singleton dimension for channels\n \n# # Forward pass for classification\n# classification_outputs = model_class(batch_images)\n# # print(classification_outputs.shape, \"before sigmoid\")\n# # print(classification_outputs)\n# # Apply sigmoid activation to the classification outputs\n \n# # Calculate binary cross-entropy loss for each class separately\n# class_loss = class_criterion(classification_outputs, batch_labels)\n \n# class_optimizer.zero_grad()\n# class_loss.backward()\n# class_optimizer.step()\n \n# classification_outputs = torch.sigmoid(classification_outputs)\n# # print(classification_outputs.shape, \"after sigmoid\")\n# # print(classification_outputs)\n \n# # Calculate accuracy and precision\n# predicted_labels = (classification_outputs > 0.5).float()\n# print(\"predicted labels\", predicted_labels)\n# all_predicted_labels.append(predicted_labels.cpu().numpy())\n# all_batch_labels.append(batch_labels.cpu().numpy())\n# y_true = batch_labels.flatten()\n# y_pred = predicted_labels.flatten()\n# # true_positives = (predicted_labels * batch_labels).sum(dim=0)\n# # false_positives = ((1 - batch_labels) * predicted_labels).sum(dim=0)\n# # false_negatives = (batch_labels * (1 - predicted_labels)).sum(dim=0)\n# # true_negatives = ((1 - batch_labels) * (1 - predicted_labels)).sum(dim=0)\n# # accuracy = (true_positives + true_negatives) / (true_positives + true_negatives + false_positives + false_negatives)\n# # precision = true_positives / (true_positives + false_positives)\n \n# accuracy = metrics.accuracy_score(y_true, y_pred)\n# print(\"Batch Classification Loss:\", class_loss.item())\n# # print(\"precision\",precision)\n# print(\"Batch accuracy\",accuracy)\n \n# # Flatten arrays for binary classification metrics\n# all_batch_labels = np.concatenate(all_batch_labels, axis=0)\n# all_predicted_labels = np.concatenate(all_predicted_labels, axis=0)\n# y_true = all_batch_labels.flatten()\n# y_pred = all_predicted_labels.flatten()\n \n\n# precision = metrics.precision_score(y_true, y_pred, average='binary')\n# accuracy = metrics.accuracy_score(y_true, y_pred)\n# print(\"Epoch\", epoch)\n# print(f'Precision: {precision:.4f}')\n# print(f'Accuracy: {accuracy:.4f}')\n \n# if epoch % 50 == 0:\n# torch.save({\n# 'epoch': epoch,\n# 'model_state_dict': model_class.state_dict(),\n# 'optimizer_state_dict': class_optimizer.state_dict(),\n# 'loss': class_loss.item()\n# # Add any other information you want to save\n# }, f'/kaggle/working/denseNet_epoch_{epoch}.pth')\n\n# # Validation loop \n# with torch.no_grad():\n# model_class.eval()\n# all_predicted_labels = []\n# all_batch_labels = []\n# for batch_images, batch_labels in val_loader:\n# batch_images = batch_images.to(torch.float32).to(device)\n# batch_labels = batch_labels.to(torch.float32).to(device)\n\n# batch_images = batch_images.unsqueeze(1)\n\n# classification_outputs = model_class(batch_images)\n# classification_outputs = torch.sigmoid(classification_outputs)\n\n# predicted_labels = (classification_outputs > 0.5).float()\n\n# all_predicted_labels.append(predicted_labels.cpu().numpy())\n# all_batch_labels.append(batch_labels.cpu().numpy())\n \n# # Flatten arrays for binary classification metrics\n# all_batch_labels = np.concatenate(all_batch_labels, axis=0)\n# all_predicted_labels = np.concatenate(all_predicted_labels, axis=0)\n\n# y_true = all_batch_labels.flatten()\n# y_pred = all_predicted_labels.flatten()\n\n# # Precision\n# precision = metrics.precision_score(y_true, y_pred, average='binary')\n\n\n# # Accuracy\n# accuracy = metrics.accuracy_score(y_true, y_pred)\n\n \n# print(\"Validation Results: \")\n# print(f'Precision: {precision:.4f}')\n# print(f'Accuracy: {accuracy:.4f}')\n# # Testing loop\n# with torch.no_grad():\n# model_class.eval()\n# all_predicted_labels = []\n# all_batch_labels = []\n# for batch_images, batch_labels in test_loader:\n# batch_images = batch_images.to(torch.float32).to(device)\n# batch_labels = batch_labels.to(torch.float32).to(device)\n\n# batch_images = batch_images.unsqueeze(1)\n\n# classification_outputs = model_class(batch_images)\n# classification_outputs = torch.sigmoid(classification_outputs)\n\n# predicted_labels = (classification_outputs > 0.5).float()\n\n# all_predicted_labels.append(predicted_labels.cpu().numpy())\n# all_batch_labels.append(batch_labels.cpu().numpy())\n \n# # Flatten arrays for binary classification metrics\n# all_batch_labels = np.concatenate(all_batch_labels, axis=0)\n# all_predicted_labels = np.concatenate(all_predicted_labels, axis=0)\n\n# y_true = all_batch_labels.flatten()\n# y_pred = all_predicted_labels.flatten()\n# print(\"y_true shape \",y_true.shape)\n# print(\"y_pred_shape\" ,y_pred.shape)\n# # Precision, Recall, F1 Score\n# precision = metrics.precision_score(y_true, y_pred, average='binary')\n# recall = metrics.recall_score(y_true, y_pred, average='binary')\n# f1_score = metrics.f1_score(y_true, y_pred, average='binary')\n\n# # Accuracy\n# accuracy = metrics.accuracy_score(y_true, y_pred)\n\n# # AUC\n# fpr, tpr, thresholds = metrics.roc_curve(y_true, y_pred)\n# auc = metrics.auc(fpr, tpr)\n \n# print(\"Testing Results: \")\n# print(f'Precision: {precision:.4f}')\n# print(f'Recall: {recall:.4f}')\n# print(f'F1 Score: {f1_score:.4f}')\n# print(f'Accuracy: {accuracy:.4f}')\n# print(f'AUC: {auc:.4f}')\n \n# # Reshape predictions for multilabel classification metrics\n# y_true_multilabel = all_batch_labels.T\n# y_pred_multilabel = all_predicted_labels.T\n# print(y_true_multilabel.shape)\n# print(y_pred_multilabel.shape)\n# y_true_1 = y_true_multilabel[:, -1]\n# y_pred_1 = y_pred_multilabel[:, -1]\n# # Compute precision, recall, F1 score, and support for each class\n# precision, recall, f1_score, support = precision_recall_fscore_support(y_true_multilabel, y_pred_multilabel, average=None)\n\n# # Print scores for each class\n# for i in range(num_classes):\n# class_accuracy = accuracy_score(y_true_multilabel[i, :], y_pred_multilabel[i, :])\n# class_precision = precision_score(y_true_multilabel[i, :], y_pred_multilabel[i, :])\n\n# print(f\" Class {i + 1}:\")\n# print(f\" Accuracy: {class_accuracy:.4f}\")\n# print(f\" Precision: {class_precision:.4f}\")\n# print()","metadata":{"execution":{"iopub.status.busy":"2024-02-20T13:30:18.105666Z","iopub.execute_input":"2024-02-20T13:30:18.106141Z","iopub.status.idle":"2024-02-20T13:30:18.808639Z","shell.execute_reply.started":"2024-02-20T13:30:18.106103Z","shell.execute_reply":"2024-02-20T13:30:18.806848Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import os\nimport pandas as pd\nimport numpy as np\nimport nibabel as nib\nimport torch\nimport torch.nn as nn\nimport torch.optim as optim\nfrom torch.utils.data import Dataset, DataLoader\nfrom torchvision import transforms\nfrom sklearn.model_selection import train_test_split\nfrom scipy.ndimage import zoom\nfrom sklearn.metrics import confusion_matrix\nfrom sklearn.metrics import accuracy_score, precision_score\n# Create a function to move data to the device\ndef move_data_to_device(data, device):\n return data.to(torch.float32).to(device)\n\nclass CustomDataset(Dataset):\n def __init__(self, image_paths, labels, transform=None):\n self.image_paths = image_paths\n self.labels = labels\n self.transform = transform\n\n def __len__(self):\n return len(self.image_paths)\n\n def __getitem__(self, idx):\n image_path = self.image_paths[idx]\n\n # Load the 3D NIfTI image using nibabel\n image = nib.load(image_path).get_fdata()\n\n # Apply transformations if provided to the image\n if self.transform:\n image = self.transform(image)\n\n label = torch.tensor(self.labels[idx], dtype=torch.float32)\n \n return image, label\n\nimport torch.nn.functional as F\n\n# Function to resize NIfTI data\ndef resize_nifti(nifti_data, target_shape):\n factors = (target_shape[0] / nifti_data.shape[0],\n target_shape[1] / nifti_data.shape[1],\n target_shape[2] / nifti_data.shape[2])\n resized_data = zoom(nifti_data, factors, order=3) # Cubic interpolation (higher quality)\n return resized_data\n\n# Paths and settings\ncsv_file = '/kaggle/input/unhealthy-csv-file/combined_data (6).csv' # Update with the correct path\nbatch_size = 32\nnum_workers = 4 # Number of CPU cores to use for data loading\nnum_classes = 14 # Number of classes\ndesired_shape = (128, 128, 128)\ndevice = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\n\n# Define transformations if needed\ntransform = transforms.Compose([\n transforms.ToTensor(), # Convert to tensor\n # Add more transformations if necessary\n])\n\n# Load the CSV file\ndata = pd.read_csv(csv_file).head(1280)\n\n# Remove the extra space from the column name\ndata.columns = data.columns.str.strip()\n\n# Assuming 'data' is your DataFrame\ndata_length = len(data)\nprint(\"Length of DataFrame:\", data_length)\n\n# Split the data into training, validation, and test sets\ntrain_data, temp_data = train_test_split(data, test_size=0.2, random_state=42)\nval_data, test_data = train_test_split(temp_data, test_size=0.5, random_state=42)\n\n# Set the display option to show all rows\npd.set_option('display.max_rows', None)\n\nindex_values = train_data.index.values\n\n# Reset the display option to its default value (if needed)\npd.reset_option('display.max_rows')\n\n# Extract file paths and labels from the data\ntrain_paths = train_data['file_path'].values\ntrain_labels = train_data[['bowel_healthy','bowel_injury','extravasation_healthy','extravasation_injury','kidney_healthy','kidney_low','kidney_high','liver_healthy','liver_low','liver_high','spleen_healthy','spleen_low','spleen_high','any_injury']].values\n\nval_paths = val_data['file_path'].values\nval_labels = val_data[['bowel_healthy','bowel_injury','extravasation_healthy','extravasation_injury','kidney_healthy','kidney_low','kidney_high','liver_healthy','liver_low','liver_high','spleen_healthy','spleen_low','spleen_high','any_injury']].values\n\ntest_paths = test_data['file_path'].values\ntest_labels = test_data[['bowel_healthy','bowel_injury','extravasation_healthy','extravasation_injury','kidney_healthy','kidney_low','kidney_high','liver_healthy','liver_low','liver_high','spleen_healthy','spleen_low','spleen_high','any_injury']].values\n\n# Instantiate the datasets\ntrain_dataset = CustomDataset(train_paths, train_labels, transform=transform)\nprint('len of train_dataset', len(train_dataset))\nval_dataset = CustomDataset(val_paths, val_labels, transform=transform)\ntest_dataset = CustomDataset(test_paths, test_labels, transform=transform)\n\n# Instantiate the data loaders\ntrain_loader = DataLoader(train_dataset, batch_size=batch_size, shuffle=True, num_workers=num_workers, drop_last=True)\nprint('train_loader', len(train_loader))\nval_loader = DataLoader(val_dataset, batch_size=batch_size, shuffle=False, num_workers=num_workers)\ntest_loader = DataLoader(test_dataset, batch_size=batch_size, shuffle=False, num_workers=num_workers)\nprint(train_loader)\n\n# Instantiate the model with the appropriate number of classes for classification\nin_channels = 1 # Input channels (e.g., for grayscale images or volumes)\nnum_classes_classification = 14 # Number of classes for classification\nmodel_class = ConvolutionalVisionTransformer(in_channels, num_classes_classification)\n\n# Count the number of parameters\ntotal_params_class = sum(p.numel() for p in model_class.parameters())\nprint(f\"Total Trainable Parameters for Classification: {total_params_class}\")\n\n# Define loss function and optimizer\nclass_criterion = nn.BCEWithLogitsLoss() # Binary Cross-Entropy loss for classification\nclass_optimizer = optim.Adam(model_class.parameters(), lr=0.001)\n\n# Training loop\n# Training loop\nclass_labels = ['bowel', 'extravasation', 'kidney', 'liver', 'spleen', 'any_injury']\n\n\n# Training loop\nnum_epochs = 100\ntrue_pos=[]\ntrue_neg=[]\nfalse_pos=[]\nfalse_neg=[]\nfor epoch in range(num_epochs):\n model_class.train()\n running_loss = 0.0\n correct_train = 0\n total_train = 0\n batch_number=0\n for batch_images, batch_labels in train_loader:\n batch_number=batch_number+1\n print(batch_number)\n # Move data to the GPU if available\n batch_images = batch_images.to(torch.float32).to(device)\n batch_labels = batch_labels.to(torch.float32).to(device)\n\n # Assuming batch_images has shape (batch_size, num_frames, num_channels, height, width)\n batch_images = batch_images.unsqueeze(1) # Add a singleton dimension for channels\n \n # Forward pass for classification\n classification_outputs = model_class(batch_images)\n \n # Apply sigmoid activation to the classification outputs\n classification_outputs = torch.sigmoid(classification_outputs)\n \n # Calculate binary cross-entropy loss for each class separately\n class_loss = class_criterion(classification_outputs, batch_labels)\n \n class_optimizer.zero_grad()\n class_loss.backward()\n class_optimizer.step()\n\n # Calculate accuracy and precision\n predicted_labels = (classification_outputs > 0.5).float()\n true_positives = (predicted_labels * batch_labels).sum(dim=0)\n true_pos.append(true_positives)\n false_positives = ((1 - batch_labels) * predicted_labels).sum(dim=0)\n false_pos.append(false_positives)\n false_negatives = (batch_labels * (1 - predicted_labels)).sum(dim=0)\n false_neg.append(false_negatives)\n true_negatives = ((1 - batch_labels) * (1 - predicted_labels)).sum(dim=0)\n accuracy = (true_positives + true_negatives) / (true_positives + true_negatives + false_positives + false_negatives)\n true_neg.append(true_negatives)\n precision = true_positives / (true_positives + false_positives)\n \n print(\"Total Classification Loss:\", class_loss.item())\n print(\"precision\",precision)\n print(\"accuracy\",accuracy)\n if epoch % 50 == 0:\n torch.save({\n 'epoch': epoch,\n 'model_state_dict': model_class.state_dict(),\n 'optimizer_state_dict': class_optimizer.state_dict(),\n 'loss': class_loss.item()\n # Add any other information you want to save\n }, f'/kaggle/working/transformerWA_epoch_{epoch}.pth')\n accuracy = (sum(true_pos) + sum(true_neg)) / (sum(true_pos) + sum(true_neg) + sum(false_pos) + sum(false_neg))\n precision = sum(true_pos)/ (sum(true_pos) + sum(false_pos))\n print(\"precision\",precision)\n print(\"accuracy\",accuracy)\n\n # Calculate evaluation metrics after all epochs\nwith torch.no_grad():\n model_class.eval()\n all_predicted_labels = []\n all_batch_labels = []\n for batch_images, batch_labels in test_loader:\n batch_images = batch_images.to(torch.float32).to(device)\n batch_labels = batch_labels.to(torch.float32).to(device)\n\n batch_images = batch_images.unsqueeze(1)\n\n classification_outputs = model_class(batch_images)\n classification_outputs = torch.sigmoid(classification_outputs)\n\n predicted_labels = (classification_outputs > 0.5).float()\n\n all_predicted_labels.append(predicted_labels.cpu().numpy())\n all_batch_labels.append(batch_labels.cpu().numpy())\n \n\n all_predicted_labels = np.concatenate(all_predicted_labels, axis=0)\n all_batch_labels = np.concatenate(all_batch_labels, axis=0)\n \n # Calculate accuracy, precision, recall, and F1 score\n true_positives = np.sum(all_predicted_labels * all_batch_labels, axis=0)\n false_positives = np.sum(all_predicted_labels * (1 - all_batch_labels), axis=0)\n false_negatives = np.sum((1 - all_predicted_labels) * all_batch_labels, axis=0)\n\n micro_precision = np.sum(true_positives) / (np.sum(true_positives) + np.sum(false_positives))\n micro_accuracy = np.mean((all_predicted_labels == all_batch_labels).all(axis=1)) # Correctly predict all labels\n micro_recall = np.sum(true_positives) / (np.sum(true_positives) + np.sum(false_negatives))\n micro_f1_score = 2 * micro_precision * micro_recall / (micro_precision + micro_recall)\n num_classes = all_batch_labels.shape[1]\n macro_precision = np.mean(true_positives / (true_positives + false_positives), axis=0)\n macro_accuracy = np.mean((all_predicted_labels == all_batch_labels).all(axis=1)) # Same as micro-accuracy\n macro_recall = np.mean(true_positives / (true_positives + false_negatives), axis=0)\n macro_f1_score = 2 * macro_precision * macro_recall / (macro_precision + macro_recall)\n\n print(\"Micro Precision:\", micro_precision)\n print(\"Micro Accuracy:\", micro_accuracy)\n print(\"Micro F1-score:\", micro_f1_score)\n print(\"Macro Precision:\", macro_precision)\n print(\"Macro Accuracy:\", macro_accuracy)\n print(\"Macro Recall:\", macro_recall)\n print(\"Macro F1-score:\", macro_f1_score)\n j=0\n for i in range(0,4,2):\n total_true_positives = 0\n total_false_positives = 0\n total_false_negatives = 0\n total_true_negatives = 0\n \n bowel_pred=predicted_labels[:,i:i+2]\n bowel_truth_label=batch_labels[:,i:i+2]\n print(bowel_pred.shape)\n print(bowel_truth_label.shape)\n true_positives = (bowel_pred * bowel_truth_label).sum(dim=0)\n false_positives = ((1 - bowel_truth_label) * bowel_pred).sum(dim=0)\n false_negatives = (bowel_truth_label * (1 - bowel_pred)).sum(dim=0)\n true_negatives = ((1 - bowel_truth_label) * (1 - bowel_pred)).sum(dim=0)\n accuracy = (true_positives + true_negatives) / (true_positives + true_negatives + false_positives + false_negatives)\n precision = true_positives / (true_positives + false_positives)\n print(class_labels[j])\n j=j+1\n total_true_positives += true_positives.sum()\n total_false_positives += false_positives.sum()\n total_false_negatives += false_negatives.sum()\n total_true_negatives += true_negatives.sum()\n\n# Calculate micro-averaged accuracy and precision\n micro_accuracy = (total_true_positives + total_true_negatives) / (total_true_positives + total_true_negatives + total_false_positives + total_false_negatives)\n micro_precision = total_true_positives / (total_true_positives + total_false_positives)\n print(\"Micro Precision:\", micro_precision)\n print(\"Micro Accuracy:\", micro_accuracy)\n \n for i in range(4,14,3):\n total_true_positives = 0\n total_false_positives = 0\n total_false_negatives = 0\n total_true_negatives = 0\n bowel_pred=predicted_labels[:,i:i+3]\n bowel_truth_label=batch_labels[:,i:i+3]\n print(bowel_pred.shape)\n print(bowel_truth_label.shape)\n true_positives = (bowel_pred * bowel_truth_label).sum(dim=0)\n false_positives = ((1 - bowel_truth_label) * bowel_pred).sum(dim=0)\n false_negatives = (bowel_truth_label * (1 - bowel_pred)).sum(dim=0)\n true_negatives = ((1 - bowel_truth_label) * (1 - bowel_pred)).sum(dim=0)\n accuracy = (true_positives + true_negatives) / (true_positives + true_negatives + false_positives + false_negatives)\n precision = true_positives / (true_positives + false_positives)\n print(class_labels[j])\n j=j+1\n total_true_positives += true_positives.sum()\n total_false_positives += false_positives.sum()\n total_false_negatives += false_negatives.sum()\n total_true_negatives += true_negatives.sum()\n\n# Calculate micro-averaged accuracy and precision\n micro_accuracy = (total_true_positives + total_true_negatives) / (total_true_positives + total_true_negatives + total_false_positives + total_false_negatives)\n micro_precision = total_true_positives / (total_true_positives + total_false_positives)\n print(\"Micro Precision:\", micro_precision)\n print(\"Micro Accuracy:\", micro_accuracy)\n ","metadata":{"execution":{"iopub.status.busy":"2024-02-16T05:36:39.038218Z","iopub.status.idle":"2024-02-16T05:36:39.038871Z","shell.execute_reply.started":"2024-02-16T05:36:39.038556Z","shell.execute_reply":"2024-02-16T05:36:39.038588Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]}]}