Canopus Walking Dead adp Lo R A

Introduction

The Canopus-Walking-Dead-adp-LoRA model is designed for generating text-to-image outputs using stable diffusion techniques with a focus on themes inspired by the "Walking Dead" franchise. It utilizes the LoRA (Low-Rank Adaptation) mechanism to enhance image generation capabilities.

Architecture

• Base Model: StabilityAI's Stable-Diffusion-XL-Base-1.0.
• Technology: Utilizes techniques such as LoRA for efficient image adaptation and generation.
• Tags: text-to-image, stable-diffusion, lora, diffusers, template:sd-lora.

Training

• Optimizer: AdamW.
• Learning Rate Scheduler: Constant.
• Noise Offset: 0.03.
• Multires Noise Discount: 0.1.
• Network Dim: 64.
• Network Alpha: 32.
• Epochs: 10.
• Repeat: 20.
• Save Every N Epochs: 1.

Guide: Running Locally

1. Setup the Pipeline:

   from diffusers import StableDiffusionXLPipeline, EulerAncestralDiscreteScheduler
   
   pipe = StableDiffusionXLPipeline.from_pretrained(
       "stabilityai/stable-diffusion-xl-base-1.0",
       torch_dtype=torch.float16,
       use_safetensors=True,
   )
   pipe.scheduler = EulerAncestralDiscreteScheduler.from_config(pipe.scheduler.config)
   pipe.load_lora_weights("Canopus-Walking-Dead-adp-LoRA", weight_name="Canopus-Walking-Dead-adp-LoRA.safetensors", adapter_name="rlms")
   pipe.set_adapters("rlms")
   pipe.to("cuda")
   

2. Trigger Prompts:

   • Use specific prompts like:
     • "A portrait of Daryl Dixon from the walking dead, action pose, blood stains, battle flames background, anime style, high resolution, <lora:walking_dead_offset:1>"
     • Use the keyword "Walking Dead" to trigger image generation.

3. Cloud GPUs: For optimal performance, consider using cloud-based GPU services such as AWS EC2, Google Cloud, or Azure.

License

This model is released under the CreativeML-OpenRAIL-M license, ensuring openness and accessibility for modification and distribution within the specified terms.