From 611dda1f2e0060af93b329ad1f196788635424b6 Mon Sep 17 00:00:00 2001 From: Dilara Gokay Date: Wed, 13 Apr 2022 23:31:56 +0200 Subject: [PATCH] Update air.ipynb (#3066) --- tutorial/source/air.ipynb | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tutorial/source/air.ipynb b/tutorial/source/air.ipynb index 0966e04558..1be74e9696 100644 --- a/tutorial/source/air.ipynb +++ b/tutorial/source/air.ipynb @@ -119,7 +119,7 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "Inference is performed in this model using [amortized stochastic variational inference](svi_part_i.ipynb) (SVI). The parameters of the neural network are also optimized during inference. Performing inference in such rich models is always difficult, but the presence of discrete choices (the number of steps in this case) makes inference in this model particularly tricky. For this reason the authors use a technique called data dependent baselines to achieve good performance. This technique can be implemented in Pyro, and we'll see how later in the tutorial.\n", + "Inference is performed in this model using [amortized stochastic variational inference](svi_part_ii.ipynb) (SVI). The parameters of the neural network are also optimized during inference. Performing inference in such rich models is always difficult, but the presence of discrete choices (the number of steps in this case) makes inference in this model particularly tricky. For this reason the authors use a technique called data dependent baselines to achieve good performance. This technique can be implemented in Pyro, and we'll see how later in the tutorial.\n", "\n", "## Model\n", "\n",