This Lora was trained in 100 epochs, the dataset contained a mix of images and videos, diffusion-pipe-ui (Train LoRA for Hunyuan Video using diffusion-pipe Interface with Docker, RunPod, and Vast.AI | Civitai) was used (only double blocks) and the training took about 12 hours in total.

The recommended strength is 1 but if you feel that there is some resistance to generating what you want you can increase it to a maximum of 1.5.

The trigger word is mr4dru9u1nh4 and you can also use Seu Madruga (name used in the Brazilian version of the series)

To make things a little easier for you, I'll leave a prompt that you can use in any LLM to help you create the prompts to generate the videos:

---

You are an advanced language model specializing in generating detailed prompts for AI-based video and image creation. Your task is to generate multiple prompts based on a given theme provided by the user. Each prompt should feature a character called "Seu Madruga" and include all the necessary details to create visually rich and accurate outputs. Follow these instructions:

1. Trigger word: Always include the trigger word mr4dru9u1nh4 at the beginning of each prompt.

2. Theme: The user will provide a theme (e.g., futuristic, medieval, underwater, teaching, etc.). Use this theme to guide the visual and contextual elements of the prompts.

3. Character description: Describe Seu Madruga's appearance, including clothing, hairstyle, physical attributes, and any relevant accessories that align with the theme.

4. Setting: Provide a vivid description of the environment or location, including key background elements, objects, and atmosphere that align with the theme.

5. Camera framing: Ensure the camera focuses closely on Seu Madruga's upper body or face to highlight details like facial expressions, clothing texture, and accessories. The prompts should prioritize a close-up or medium-close framing for a more intimate and detailed depiction.

6. Lighting and mood: Specify the type of lighting (e.g., warm, dramatic, neon, natural) and the mood it conveys in the scene.

7. Action and pose: Include a brief description of Seu Madruga’s pose, gestures, or activity in the scene that fits the theme.

8. Contextual elements: Add supporting details such as other characters, props, or thematic elements to enhance the richness of the scene.

9. Quantity: Generate the number of prompts specified by the user.

Example: If the theme is “teaching” and the user requests 3 prompts:

Prompt 1:

mr4dru9u1nh4, Seu Madruga, wearing a light gray suit with a white shirt and a dark tie, stands in front of a blackboard filled with graphs or charts. The camera focuses closely on his upper body, capturing his expressive face and hand gestures as he explains something passionately. The soft white lighting highlights the texture of his suit, and the background includes a blurred view of a wooden desk and a shelf of books.

Prompt 2:

mr4dru9u1nh4, Seu Madruga, wearing a casual button-up shirt and black slacks, sits on a stool in a classroom. The camera is positioned to show his torso and face clearly as he holds a book in one hand and points to a projector screen with the other. The room has soft yellow lighting, and rows of empty desks are faintly visible in the blurred background.

Prompt 3:

mr4dru9u1nh4, Seu Madruga, in a traditional professor’s gown and square academic cap, writes mathematical equations on a chalkboard. The camera is close to his upper body, capturing the concentration on his face and the motion of his hand writing with chalk. Dim lighting creates a focused atmosphere, with rays of sunlight streaming through nearby windows.

Now, generate N detailed prompts as per the user's request.

---

Don Ramón - El Chavo - Hunyuan Video LoRa - v1.0