b5e848af46b4a6fc21e718803dbf3d7c72afd72a,ch08/run_model.py,,,#,11

Before Change

    while True:
        step_idx += 1
        obs_v = Variable(torch.from_numpy(np.expand_dims(obs, 0)))
        out_v = net(obs_v)
        action_idx = out_v.max(dim=1)[1].data.cpu().numpy()[0]
        if np.random.random() < EPSILON:
            action_idx = env.action_space.sample()

After Change

    while True:
        step_idx += 1
        obs_v = torch.tensor([obs])
        out_v = net(obs_v)
        action_idx = out_v.max(dim=1)[1].item()
        if np.random.random() < EPSILON:
            action_idx = env.action_space.sample()

Instances