diff --git a/dqn_zoo/iqn/agent.py b/dqn_zoo/iqn/agent.py
index e1dc997..80ea8fb 100644
--- a/dqn_zoo/iqn/agent.py
+++ b/dqn_zoo/iqn/agent.py
@@ -76,7 +76,9 @@ def select_action(rng_key, network_params, s_t):
       tau_t = _sample_tau(tau_key, (1, tau_samples))
       q_t = network.apply(
           network_params, apply_key, IqnInputs(s_t[None, ...], tau_t)
-      ).q_values[0]
+      ).q_values[
+          0
+      ]  # pytype: disable=wrong-arg-types  # jax-ndarray
       a_t = distrax.EpsilonGreedy(q_t, exploration_epsilon).sample(
           seed=policy_key
       )
@@ -188,15 +190,15 @@ def loss_fn(online_params, target_params, transitions, rng_key):
       _, *apply_keys = jax.random.split(rng_key, 4)
       dist_q_tm1 = network.apply(
           online_params, apply_keys[0], IqnInputs(transitions.s_tm1, tau_tm1)
-      ).q_dist
+      ).q_dist  # pytype: disable=wrong-arg-types  # jax-ndarray
       dist_q_t_selector = network.apply(
           target_params,
           apply_keys[1],
           IqnInputs(transitions.s_t, tau_t_selector),
-      ).q_dist
+      ).q_dist  # pytype: disable=wrong-arg-types  # jax-ndarray
       dist_q_target_t = network.apply(
           target_params, apply_keys[2], IqnInputs(transitions.s_t, tau_t)
-      ).q_dist
+      ).q_dist  # pytype: disable=wrong-arg-types  # jax-ndarray
       losses = _batch_quantile_q_learning(
           dist_q_tm1,
           tau_tm1,
@@ -229,7 +231,9 @@ def select_action(rng_key, network_params, s_t, exploration_epsilon):
       tau_t = _sample_tau(sample_key, (1, tau_samples_policy))
       q_t = network.apply(
           network_params, apply_key, IqnInputs(s_t[None, ...], tau_t)
-      ).q_values[0]
+      ).q_values[
+          0
+      ]  # pytype: disable=wrong-arg-types  # jax-ndarray
       a_t = distrax.EpsilonGreedy(q_t, exploration_epsilon).sample(
           seed=policy_key
       )