running code on Google Colab

[ramishni]

I appreciate any help
reading the book to:
-Chapter 4: Reinforcement Learning in the Real World – Building Cryptocurrency Trading Agents
-Training a cryptocurrency trading bot using RL

and run flowing code : Tensorflow-2-Reinforcement-Learning-Cookbook/Chapter04/sac_agent.ipynb
I got problem this part:
def test(self, render=True, fps=30, filename="test_render.mp4"):
cur_state, done, rewards = self.env.reset(), False, 0
# video = imageio.get_writer(filename, fps=fps )
video = imageio.get_writer(filename)
while not done:
action = self.act(cur_state, test=True)
next_state, reward, done, _ = self.env.step(action[0])
cur_state = next_state
rewards += reward
if render:
video.append_data(self.env.render(mode="rgb_array"))
video.close()
return rewards
giving this error

ValueError Traceback (most recent call last)
in ()
11 # sac.train(max_epochs=1000, random_epochs=100, save_freq=50)
12
---> 13 reward = sac.test()
14 print(reward)
15 # tensorboard –-logdir=logs

---

in test(self, render, fps, filename)
320 rewards += reward
321 if render:
--> 322 video.append_data(self.env.render(mode="rgb_array"))
323 video.close()
324 return rewards

/usr/local/lib/python3.7/dist-packages/imageio/core/format.py in append_data(self, im, meta)
483 # Check image data
484 if not isinstance(im, np.ndarray):
--> 485 raise ValueError("append_data requires ndarray as first arg")
486 # Get total meta dict
487 total_meta = {}

ValueError: append_data requires ndarray as first arg