---
tags:
- text-to-image
- stable-diffusion
- lora
- diffusers
- template:sd-lora
widget:
- text: '-'
  output:
    url: images/result_juggernaut_tscd_8step.png
base_model: digiplay/Juggernaut_final
instance_prompt: null
license: mit
---
# TSCD-Juggernaut_final

<Gallery />

## Model description 

## Model description 

Fine-tune a distill LoRA version of [Lykon&amp;#x2F;dreamshaper-7](https:&amp;#x2F;&amp;#x2F;huggingface.co&amp;#x2F;Lykon&amp;#x2F;dreamshaper-7) using Trajectory Segmented Consistency Model (TSCD) introduce in [Hyper-SD: Trajectory Segmented Consistency Model for
Efficient Image Synthesis](https:&amp;#x2F;&amp;#x2F;arxiv.org&amp;#x2F;abs&amp;#x2F;2404.13686)

This LoRA is fine-tuned on [laion&amp;#x2F;conceptual-captions-12m-webdataset](https:&amp;#x2F;&amp;#x2F;huggingface.co&amp;#x2F;datasets&amp;#x2F;laion&amp;#x2F;conceptual-captions-12m-webdataset) for 16.000 iterations using unofficial training implementation at [https:&amp;#x2F;&amp;#x2F;github.com&amp;#x2F;rootonchair&amp;#x2F;consistency_models_distill](https:&#x2F;&#x2F;github.com&#x2F;rootonchair&#x2F;consistency_models_distill)

# Usage


## Download model

Weights for this model are available in Safetensors format.

[Download](/rootonchair/tscd_juggernaut_final/tree/main) them in the Files & versions tab.