Behavioral control meeting

OpenAI GPT-3

Description of why GPT-3 is cool https://towardsdatascience.com/gpt-3-creative-potential-of-nlp-d5ccae16c1ab Requesting OpenAI GPT-3 access https://forms.office.com/Pages/ResponsePage.aspx?id=VsqMpNrmTkioFJyEllK8sx3ELsv0PEhHphhNz30FttVUNkYwTlNPMVI1V0lXNjExMlExUlc4SE5YSS4u

SAVC2_run.py

Lines 165-184

  while env.day <= 60:
        step = env.day
        day_of_week = env.day % 7
        print("Day: " + str(step))
        if not start_flag:
            action = env.action_space.sample()
            next_state, reward, done, info = env.step(action)
            state = np.copy(next_state)
            start_flag = True
            continue

        if env.day <= 30:
            action = env.action_space.sample()  # Sample random action
            next_state, reward, done, info = env.step(action)

            memory.push((state, action, reward, next_state, done))

            state = np.copy(next_state)
            continue

Why is there the restriction on 30?

How can we produce the out of office indicators from the reward? Where is that code? Would we need to do that separately?

---

uid: 202007161700 tags: #meetings #raise

---

---

---