Created using Colaboratory
This commit is contained in:
@@ -4,7 +4,6 @@
|
|||||||
"metadata": {
|
"metadata": {
|
||||||
"colab": {
|
"colab": {
|
||||||
"provenance": [],
|
"provenance": [],
|
||||||
"authorship_tag": "ABX9TyNFsCOnucz1nQt7PBEnKeTV",
|
|
||||||
"include_colab_link": true
|
"include_colab_link": true
|
||||||
},
|
},
|
||||||
"kernelspec": {
|
"kernelspec": {
|
||||||
@@ -169,7 +168,7 @@
|
|||||||
{
|
{
|
||||||
"cell_type": "markdown",
|
"cell_type": "markdown",
|
||||||
"source": [
|
"source": [
|
||||||
"Because the function changes much faster in $\\phi_1$ than in $\\phi_0$, there is no great step size to choose. If we set the step size so that it makes sensible progress in the $\\phi_1$ direction, then it takes many iterations to converge. If we set the step size so that we make sensible progress in the $\\phi_{0}$ direction, then the path oscillates in the $\\phi_1$ direction. \n",
|
"Because the function changes much faster in $\\phi_1$ than in $\\phi_0$, there is no great step size to choose. If we set the step size so that it makes sensible progress in the $\\phi_1$ direction, then it takes many iterations to converge. If we set the step size so that we make sensible progress in the $\\phi_0$ direction, then the path oscillates in the $\\phi_1$ direction. \n",
|
||||||
"\n",
|
"\n",
|
||||||
"This motivates Adam. At the core of Adam is the idea that we should just determine which way is downhill along each axis (i.e. left/right for $\\phi_0$ or up/down for $\\phi_1$) and move a fixed distance in that direction."
|
"This motivates Adam. At the core of Adam is the idea that we should just determine which way is downhill along each axis (i.e. left/right for $\\phi_0$ or up/down for $\\phi_1$) and move a fixed distance in that direction."
|
||||||
],
|
],
|
||||||
|
|||||||
Reference in New Issue
Block a user