From ab2a0fc660eb37a395fef08ec5f7f0f3533c6ca8 Mon Sep 17 00:00:00 2001
From: zhangmozhe <zhangboknight@gmail.com>
Date: Mon, 21 Dec 2020 23:23:21 +0800
Subject: [PATCH] update readme

---
 warping.py | 30 ------------------------------
 1 file changed, 30 deletions(-)
 delete mode 100644 warping.py

diff --git a/warping.py b/warping.py
deleted file mode 100644
index 8cdda79..0000000
--- a/warping.py
+++ /dev/null
@@ -1,30 +0,0 @@
-import torch
-import torch.nn as nn
-import torch.nn.functional as F
-
-
-def get_grid(x):
-    torchHorizontal = (
-        torch.linspace(-1.0, 1.0, x.size(3)).view(1, 1, 1, x.size(3)).expand(x.size(0), 1, x.size(2), x.size(3))
-    )
-    torchVertical = (
-        torch.linspace(-1.0, 1.0, x.size(2)).view(1, 1, x.size(2), 1).expand(x.size(0), 1, x.size(2), x.size(3))
-    )
-    return torch.cat([torchHorizontal, torchVertical], 1).cuda()
-
-
-class WarpingLayer(nn.Module):
-    def __init__(self, device):
-        super(WarpingLayer, self).__init__()
-        self.device = device
-
-    def forward(self, x, flow):
-        # WarpingLayer uses F.VGG19_pytorch, which expects normalized grid
-        # we still output unnormalized flow for the convenience of comparing EPEs with FlowNet2 and original code
-        # so here we need to denormalize the flow
-        flow_for_grip = torch.zeros_like(flow).cuda()
-        flow_for_grip[:, 0, :, :] = flow[:, 0, :, :] / ((flow.size(3) - 1.0) / 2.0)
-        flow_for_grip[:, 1, :, :] = flow[:, 1, :, :] / ((flow.size(2) - 1.0) / 2.0)
-
-        grid = (get_grid(x) + flow_for_grip).permute(0, 2, 3, 1)
-        return F.VGG19_pytorch(x, grid, align_corners=True)