diff --git a/Chapter1_Introduction/Ch1_Introduction_TFP.ipynb b/Chapter1_Introduction/Ch1_Introduction_TFP.ipynb
index 07aede9a..9e67b92c 100644
--- a/Chapter1_Introduction/Ch1_Introduction_TFP.ipynb
+++ b/Chapter1_Introduction/Ch1_Introduction_TFP.ipynb
@@ -1208,7 +1208,7 @@
         "# Set the chain's start state.\n",
         "initial_chain_state = [\n",
         "    tf.cast(tf.reduce_mean(count_data), tf.float32) * tf.ones([], dtype=tf.float32, name=\"init_lambda1\"),\n",
-        "    tf.cast(tf.reduce_mean(count_data), tf.float32) * tf.ones([], dtype=tf.float32, name=\"init_lambda2\", tf.float32),\n",
+        "    tf.cast(tf.reduce_mean(count_data), tf.float32) * tf.ones([], dtype=tf.float32, name=\"init_lambda2\"),\n",
         "    0.5 * tf.ones([], dtype=tf.float32, name=\"init_tau\"),\n",
         "]\n",
         "\n",
@@ -1234,7 +1234,7 @@
         "\n",
         "    lambda_ = tf.gather(\n",
         "         [lambda_1, lambda_2],\n",
-        "         indices=tf.to_int32(tau * tf.cast(tf.size(count_data), tf.float32) <= tf.cast(tf.range(tf.size(count_data)), tf.float32)))\n",
+        "         indices=tf.cast(tau * tf.cast(tf.size(count_data), tf.float32) <= tf.cast(tf.range(tf.size(count_data)), tf.float32), tf.int32))\n",
         "    rv_observation = tfd.Poisson(rate=lambda_)\n",
         " \n",
         "    return (\n",
@@ -1273,11 +1273,11 @@
         "            target_log_prob_fn=unnormalized_log_posterior,\n",
         "            num_leapfrog_steps=2,\n",
         "            step_size=step_size,\n",
-        "            step_size_update_fn=tfp.mcmc.make_simple_step_size_update_policy(),\n",
+        "            step_size_update_fn=tfp.mcmc.make_simple_step_size_update_policy(num_adaptation_steps=None),\n",
         "            state_gradients_are_stopped=True),\n",
         "        bijector=unconstraining_bijectors))\n",
         "\n",
-        "tau_samples = tf.floor(posterior_tau * tf.cast(tf.size(count_data)), tf.float32)\n",
+        "tau_samples = tf.floor(posterior_tau * tf.cast(tf.size(count_data), dtype=tf.float32))\n",
         "\n",
         "# tau_samples, lambda_1_samples, lambda_2_samples contain\n",
         "# N samples from the corresponding posterior distribution\n",
diff --git a/Chapter3_MCMC/Ch3_IntroMCMC_TFP.ipynb b/Chapter3_MCMC/Ch3_IntroMCMC_TFP.ipynb
index c1257def..5ccdf3df 100644
--- a/Chapter3_MCMC/Ch3_IntroMCMC_TFP.ipynb
+++ b/Chapter3_MCMC/Ch3_IntroMCMC_TFP.ipynb
@@ -461,8 +461,12 @@
     "\n",
     "# plotting details.\n",
     "x_ = y_ = np.linspace(.01, 5, 100)\n",
-    "likelihood_x_ = evaluate(tfd.Poisson(rate=x_).prob(data_[:, 0]))\n",
-    "likelihood_y_ = evaluate(tfd.Poisson(rate=y_).prob(data_[:, 1]))\n",
+	"likelihood_x = tfd.Poisson(rate=tf.expand_dims(x_, 1)).prob(data_[:, 0])\n",
+	"likelihood_x = tf.reduce_prod(likelihood_x, axis=1)\n",
+    "likelihood_x_ = evaluate(likelihood_x)\n",
+	"likelihood_y = tfd.Poisson(rate=tf.expand_dims(y_, 1)).prob(data_[:, 1])\n",
+	"likelihood_y = tf.reduce_prod(likelihood_y, axis=1)\n",
+    "likelihood_y_ = evaluate(likelihood_y)\n",
     "L_ = evaluate(tf.matmul(tf.expand_dims(likelihood_x_, 1), \n",
     "                        tf.expand_dims(likelihood_y_, 0)))"
    ]