From 679910496e5244409fb3d4da85cef19af8306545 Mon Sep 17 00:00:00 2001 From: Philipp Moritz Date: Wed, 24 May 2017 18:39:37 -0700 Subject: [PATCH] fix policy gradients for mujoco domains (#589) --- examples/policy_gradient/reinforce/distributions.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/examples/policy_gradient/reinforce/distributions.py b/examples/policy_gradient/reinforce/distributions.py index 20e6f1741..2fc953a58 100644 --- a/examples/policy_gradient/reinforce/distributions.py +++ b/examples/policy_gradient/reinforce/distributions.py @@ -42,7 +42,7 @@ class Categorical(object): class DiagGaussian(object): def __init__(self, flat): self.flat = flat - mean, logstd = tf.split(1, 2, flat) + mean, logstd = tf.split(flat, 2, axis=1) self.mean = mean self.logstd = logstd self.std = tf.exp(logstd)