Creating comic strip prompts and captions for “Yoshimi Battles the Pink Robots”

Loading data from Talking to Chatbots Dataset reddgr/talking-to-chatbots-chats

😎 Help me compose a comic strip for the story of “Yoshimi Battles the Pink Robots”, the song by The Flaming Lips I will paste below. Can you write 4 prompts for Stable Diffusion images, each being a panel of the comic strip? Make sure that the prompts are good for Stable Diffusion images and capture the drawing style of a popular comic book franchise such as Marvel or DC (this is normally included at the end of the prompt). Make sure that the 4 prompts are relatively similar to each other so the image styles are not too inconsistent. I will adjust all prompts and may play around with the seeds or some inpainting features before finalizing the works, but I would like the prompts provided to save as much time as possible.
Can you also write captions for each panel?



Her name is Yoshimi
she’s a black belt in karate
working for the city
she has to discipline her body

‘Cause she knows that
it’s demanding
to defeat those evil machines
I know she can beat them

Oh Yoshimi, they don’t believe me
but you won’t let those robots eat me
Yoshimi, they don’t believe me
but you won’t let those robots defeat me

Those evil-natured robots
they’re programmed to destroy us
she’s gotta be strong to fight them
so she’s taking lots of vitamins

‘Cause she knows that
it’d be tragic
if those evil robots win
I know she can beat them

Oh Yoshimi, they don’t believe me
but you won’t let those robots defeat me
Yoshimi, they don’t believe me
but you won’t let those robots eat me

Yoshimi

‘Cause she knows that
it’d be tragic
if those evil robots win
I know she can beat them

Oh Yoshimi, they don’t believe me
but you won’t let those robots defeat me
Yoshimi, they don’t believe me
but you won’t let those robots defeat me

Oh Yoshimi, they don’t believe me
but you won’t let those robots eat me
Yoshimi, they don’t believe me
but you won’t let those robots eat me
🤖 Certainly! Here are four prompts for Stable Diffusion images along with captions for each panel, capturing the drawing style of a popular comic book franchise like Marvel or DC:

**Panel 1:**
Prompt: “Create a dynamic comic panel featuring Yoshimi, a strong and determined martial artist. She stands in a powerful karate pose, ready to take on the evil robots. The background shows a futuristic cityscape in turmoil, with menacing robots approaching.”
Caption: “Yoshimi, the Karate Defender, prepares to face the robotic menace in the heart of the city.”

**Panel 2:**
Prompt: “Illustrate a close-up of Yoshimi’s determined face as she delivers a high-kick to one of the evil robots. The impact creates sparks, showcasing her incredible strength. The robotic foe reels back from the force.”
Caption: “With a swift kick, Yoshimi strikes fear into the heart of the mechanical beast.”

**Panel 3:**
Prompt: “Show Yoshimi surrounded by a swarm of evil robots, each with menacing features. She skillfully dodges their attacks, displaying her agility and combat prowess. The scene is filled with energy and action.”
Caption: “Yoshimi dances through a storm of relentless robots, a whirlwind of skill and determination.”

**Panel 4:**
Prompt: “Create a climactic panel where Yoshimi stands triumphant, surrounded by defeated and broken robots. The city skyline in the background is clearing as the threat subsides.”
Caption: “Yoshimi emerges victorious, a symbol of bravery and resilience against the mechanical tide.”

Feel free to adjust these prompts as needed, and I hope they help bring your comic strip to life!
😎 Modify panel 2 so it doesn’t show a close up and it shows a long shot of the fight scene instead. I want to create a picture without detailed faces as I will use a random seed in each text-to-image creation. With random seeds in each creation, it’s virtually impossible to create a face that looks the same perso in two images. Also, instructional verbs such as “create”, “illustrate”, etc. are not needed in text-to-image prompts
🤖 Of course, here’s the modified prompt for Panel 2:

**Panel 2:**
Prompt: “Show Yoshimi engaged in a fierce battle with an evil robot. She delivers a high-kick with incredible force, sending the robot reeling. The background showcases a chaotic cityscape as the action unfolds.”
Caption: “Yoshimi’s powerful kick sends a robot tumbling, the cityscape bearing witness to her strength.”

This adjusted prompt should work well for creating a dynamic fight scene without focusing on detailed faces.

Hugging Face Dataset Metrics

All the conversation prompts, responses, and metrics are available to download and explore on Hugging Face dataset reddgr/talking-to-chatbots-chats: