In this tutorial, we will explore how to create our own comic using ChatGPT and Dolly 3. If you are new to these tools, this demo will showcase their power. By combining the capabilities of AI, we can achieve impressive results in less than two minutes.
To get started, you will need to log into ChatGPT. The free version is sufficient for this demonstration. Additionally, you will need to log into Bing Image Creator, which utilizes Dolly 3. A Microsoft account is required, but it provides free credits for experimentation.
Next, we will examine the prompt templates. The first template involves ChatGPT acting as a creative comic writer. It generates a four-picture comic strip with three to five-word scene descriptions. We will choose the adventure theme for this example. Let’s imagine a silly pickle going to school for the first time. ChatGPT will generate four text bubbles, one for each scene.
Once we have the template, we can copy it into ChatGPT and let it generate the scenes. Afterward, we copy the generated content back into our notepad. We will then take pieces from this content and incorporate them into prompt template 2, which will be used with Dolly 3.
Template 2 involves creating a four-picture comic strip based on a theme. We will use the silly pickle going to school for the first time as our theme. We copy the theme into the template and then copy the scene descriptions from ChatGPT into the corresponding scenes in the template. Additionally, we can choose a style for the comic strip. For this example, let’s go with a cyberpunk style.
Once we have completed all the scenes, we can proceed to Dolly 3. However, Dolly 3 has a character limit of 400. Therefore, we need to limit the content from ChatGPT to 400 characters. Afterward, we can input the content into Dolly 3 and see what it generates.
After a short wait, we are presented with our comic strip. The graphics look amazing, although it seems that there is no text in the bottom left. Overall, it is an impressive result. However, there may be some inconsistencies between the text and graphics. Feel free to experiment with different prompts and have fun with it!
