Initial commit

convolve.py

+#!/usr/bin/env python
+
+import numpy as np
+import time
+
+# Copied from https://github.com/wiseodd/hipsternet
+# Shouldn't need to directly use this
+def _get_im2col_indices(x_shape, field_height, field_width, padding=0, stride=1):
+  # First figure out what the size of the output should be
+  N, C, H, W = x_shape
+  assert (H + 2 * padding - field_height) % stride == 0
+  assert (W + 2 * padding - field_height) % stride == 0
+  out_height = (H + 2 * padding - field_height) / stride + 1
+  out_width = (W + 2 * padding - field_width) / stride + 1
+
+  i0 = np.repeat(np.arange(field_height), field_width)
+  i0 = np.tile(i0, C)
+  i1 = stride * np.repeat(np.arange(out_height), out_width)
+  j0 = np.tile(np.arange(field_width), field_height * C)
+  j1 = stride * np.tile(np.arange(out_width), out_height)
+  i = i0.reshape(-1, 1) + i1.reshape(1, -1)
+  j = j0.reshape(-1, 1) + j1.reshape(1, -1)
+
+  k = np.repeat(np.arange(C), field_height * field_width).reshape(-1, 1)
+
+  return (k, i, j)
+
+# Copied from https://github.com/wiseodd/hipsternet
+# Shouldn't need to directly use this
+def _im2col_indices(x, field_height, field_width, padding=0, stride=1):
+  """ An implementation of im2col based on some fancy indexing """
+  # Zero-pad the input
+  p = padding
+  x_padded = np.pad(x, ((0, 0), (0, 0), (p, p), (p, p)), mode='constant')
+
+  k, i, j = _get_im2col_indices(x.shape, field_height, field_width, padding,
+                               stride)
+
+  cols = x_padded[:, k, i, j]
+  C = x.shape[1]
+  cols = cols.transpose(1, 2, 0).reshape(field_height * field_width * C, -1)
+  return cols
+
+# Adapted from https://github.com/wiseodd/hipsternet
+# X is an image with dimensions (X_height, X_width)
+# W is a filter with dimensions (W_channels, W_height, W_width)
+# stride is the amount of pixels the filter is slid
+# padding is the number of border pixels that are added to the image
+# Returns the result of the convolution with dimensions (out_channels, out_height, out_width)
+def convolve(X, W, stride=1, padding=0):
+    X = X[np.newaxis, np.newaxis, :, :]
+    W = W[:,np.newaxis,:,:]
+    n_filters, d_filter, h_filter, w_filter = W.shape
+    n_x, d_x, h_x, w_x = X.shape
+    h_out = (h_x - h_filter + 2 * padding) / stride + 1
+    w_out = (w_x - w_filter + 2 * padding) / stride + 1
+
+    #if not h_out.is_integer() or not w_out.is_integer():
+    #    raise Exception('Invalid output dimension!')
+
+    h_out, w_out = int(h_out), int(w_out)
+
+    X_col = _im2col_indices(X, h_filter, w_filter, padding=padding, stride=stride)
+    W_col = W.reshape(n_filters, -1)
+
+    #print W_col.shape
+    #print X_col.shape
+    #out = W_col @ X_col 
+    out = np.dot(W_col, X_col)
+    out = out.reshape(n_filters, h_out, w_out, n_x)
+    out = out.transpose(3, 0, 1, 2)
+
+    return out.squeeze(axis=0)
+    
+if __name__ == '__main__':
+  #X = np.array([[10,11,12,13],[14,15,16,17],[18,19,20,21],[22,23,24,25]])
+  #X = np.random.rand(80,120)
+  #X = X[np.newaxis, np.newaxis, :, :]
+  #W = np.array([[[1,1],[1,1]],[[2,2],[2,2]]])
+  #W = np.random.rand(100,80,20)
+  #W = W[:,np.newaxis,:,:]
+  
+  X = np.array([[10,11,12,13],[14,15,16,17],[18,19,20,21],[22,23,24,25]])
+  W = np.array([[[1,2],[3,4]],[[2,2],[2,2]]])
+  
+  print 'X:'
+  print X
+  print 'W:'
+  print W
+  print 'Res:'
+  start = time.time()
+  out = convolve(X,W, padding=0)
+  print out
+  print time.time()-start
+