Final pre-print

GANs-in-Action · Jun 4, 2019 · 256a71f · 256a71f
1 parent ee15c43
commit 256a71f
Showing 1 changed file with 2 additions and 2 deletions.
diff --git a/chapter-6/Chapter_6_PGGAN.ipynb b/chapter-6/Chapter_6_PGGAN.ipynb
@@ -190,8 +190,8 @@
     "def equalize_learning_rate(shape, gain, fan_in=None):\n",
     "    '''\n",
     "    This adjusts the weights of every layer by the constant from He's initializer so that we adjust for the variance in the dynamic range in different features\n",
-    "    shape   :   shape of tensor (layer): these are the layer weights that depend on your implementation. \n",
-    "    The clearest (PyTorch) implementation is in our opinion here: https://github.com/facebookresearch/pytorch_GAN_zoo/blob/87d90f61cbdc19233e357f36ac342e08453736fd/models/networks/custom_layers.py#L28\n",
+    "    shape   :   shape of tensor (layer): these are the dimensions of each layer.\n",
+    "    For example, [4,4,48,3]. In this case, [kernel_size, kernel_size, number_of_filters, feature_maps]. But this will depend slightly on your implementation.\n",
     "    gain    :   typically sqrt(2)\n",
     "    fan_in  :   adjustment for the number of incoming connections as per Xavier's / He's initialization \n",
     "    '''\n",