bdbdd9b6ebb7a188a331d8e7a6911c4e6ee99b7a,tests/agents_tests/test_categorical_dqn.py,,make_distrib_ff_q_func,#Any#,142

Before Change

    n_atoms = 51
    v_max = 10
    v_min = -10
    delta_z = (v_max - v_min) / float(n_atoms - 1)
    z_values = np.array(
        [v_min + i * delta_z for i in range(n_atoms)], dtype=np.float32)
    return chainerrl.q_functions.DistributionalFCStateQFunctionWithDiscreteAction(  // NOQA
        env.observation_space.low.size, env.action_space.n,
        n_atoms=n_atoms,
        z_values=z_values,

After Change

    n_atoms = 51
    v_max = 10
    v_min = -10
    z_values = np.linspace(v_min, v_max, num=n_atoms, dtype=np.float32)
    return chainerrl.q_functions.DistributionalFCStateQFunctionWithDiscreteAction(  // NOQA
        env.observation_space.low.size, env.action_space.n,
        n_atoms=n_atoms,
        z_values=z_values,

Instances