---
title: flux-ghibsky-illustration
canonical_url: "https://www.modelscope.cn/models/AI-ModelScope/flux-ghibsky-illustration"
md_url: "https://www.modelscope.cn/models/AI-ModelScope/flux-ghibsky-illustration.md"
repository: AI-ModelScope/flux-ghibsky-illustration
last_updated: 2026-08-13
license: other
pipeline_tag: text-to-image-synthesis
tasks:
  - text-to-image-synthesis
base_model:
  - black-forest-labs/FLUX.1-dev
base_model_relation: adapter
parameters: 171.8M
tensor_type:
  - F16
library_name:
  - lora
  - diffusers
  - safetensors
  - pytorch
frameworks:
  - pytorch
downloads: 659
stars: 0
tags:
  - text-to-image
  - diffusers
  - lora
  - "template:sd-lora"
  - image-generation
  - flux
  - replicate
---

# flux-ghibsky-illustration

> flux-ghibsky-illustration - AI-ModelScope 在 ModelScope 开源的模型。🎨 Flux Ghibsky Illustration

AI-ModelScope/flux-ghibsky-illustration 是 ModelScope 魔搭社区上的 171.8M 参数text-to-image-synthesis模型，采用 other 许可，基于 black-forest-labs/FLUX.1-dev 构建。

- **Repository**: AI-ModelScope/flux-ghibsky-illustration
- **License**: other
- **Tasks**: text-to-image-synthesis
- **Parameters**: 171.8M
- **Base model**: black-forest-labs/FLUX.1-dev
- **Tags**: text-to-image, diffusers, lora, template:sd-lora, image-generation, flux, replicate
- **Downloads**: 659
- **Stars**: 0
- **Last updated**: 2026-08-13

Source: https://www.modelscope.cn/models/AI-ModelScope/flux-ghibsky-illustration

---

# 🎨 Flux Ghibsky Illustration

<Gallery />

## 🌟 Overview

The **Flux Ghibsky Illustration** model creates dreamy, enchanting landscapes by combining the whimsical charm of **Studio Ghibli** with the detailed, atmospheric beauty of **Makoto Shinkai**. 

- 🚀 **Try online:** [Run on Replicate](https://replicate.com/aleksa-codes/flux-ghibsky-illustration)
- 💾 **Download:** Get the `.safetensors` file in the [Files & Versions tab](https://huggingface.co/aleksa-codes/flux-ghibsky-illustration/tree/main)

---

## 💡 Prompting Guide

* **Trigger Word:** Always start your prompt with `GHIBSKY style`.
* **Pro Tip:** If your generated images look too realistic, add `painting` to your prompt (e.g., `GHIBSKY style painting`).

**Example Prompt:**
> `GHIBSKY style, a cozy mountain cabin covered in snow, with smoke curling from the chimney under a starry night sky`

---

## ⚙️ Training Details

| Setting | Value |
| :--- | :--- |
| **Dataset Size** | 35 images |
| **Trigger Word** | `GHIBSKY` |
| **Trainers Used** | [fal.ai](https://fal.ai/models/fal-ai/flux-lora-fast-training) & [Replicate Trainer](https://replicate.com/ostris/flux-dev-lora-trainer/train) |
| **Training Steps** | 1000 |
| **Learning Rate** | 0.0004 |
| **LoRA Rank** | 16 |
| **Batch Size** | 1 |

---

## 🛠️ Related Tools

* **[UnYellowGPT](https://unyellowgpt.com/):** Easily remove yellow or sepia color casts from AI-generated images to restore true white balance.
* **[GPT Image Captioner](https://gptcaptioner.aleksa.codes/):** An open-source tool for generating high-quality `.txt` captions for your LoRA training datasets (supports OpenAI & local Ollama).

---

## ☕ Support My Work

If you enjoy my models and tools, feel free to [buy me a coffee on Ko-fi](https://ko-fi.com/aleksacodes)!

---

## 📜 License

This model is subject to the [FLUX.1-dev License Terms](https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md).