---
tags:
- stable-diffusion
- text-to-image
- lora
- character
- diffusionpipe
---
# Character LoRA Collection: Martin, Demetri, & Isaac
This repository contains a collection of three distinct character LoRAs. Each model was trained on 30 images generated by an LLM, using the **WAN2.1 14B** I2V model as a base.
Since all models are hosted in a single repository, please follow the usage instructions carefully to load the specific character you want.
## Character LoRAs Overview
### 1. Demetri
| Filename | Trigger Words |
| ------------------------ | --------------------------------- |
| `demetri_15.safetensors` | `#demerti, running and laughing` |
### 2. Martin
| Filename | Trigger Words |
| ---------------------- | ------------------------------ |
| `martin_15.safetensors`| `#martin, 1boy` |
### 3. Isaac
| Filename | Trigger Words |
| ---------------------- | ----------------------- |
| `isaac_15.safetensors` | `#isaac, laughing` |
---
## How to Use with
Because this repository contains multiple LoRA files, you **must specify the `weight_name`** (the filename) you wish to load.
- Download the models you want and load into comfyUI
## Training Details
Base Model: The LoRAs were trained on top of the WAN2.1 14B parameter T2V model.
Training Library: The training was implemented using the diffusionpipe library.
Hyperparameters: For a detailed look at the training configuration and hyperparameters, please visit the training repository here.
## Dataset
- Each LoRA was trained on a dataset of 30 images that were generated using a Large Language Model to ensure a consistent character style.
- Sample Data Used for Training:
(This image is an example from the dataset to show the style and quality.)
## Demo Video