---
title: Ernie-Image-Flex-V1
canonical_url: "https://www.modelscope.cn/models/wikeeyang/Ernie-Image-Flex-V1"
md_url: "https://www.modelscope.cn/models/wikeeyang/Ernie-Image-Flex-V1.md"
repository: wikeeyang/Ernie-Image-Flex-V1
chinese_name: Ernie-Image-Flex-V1
last_updated: 2026-05-12
license: apache-2.0
pipeline_tag: text-to-image-synthesis
tasks:
  - text-to-image-synthesis
base_model:
  - PaddlePaddle/ERNIE-Image
base_model_relation: finetune
library_name:
  - safetensors
  - gguf
  - pytorch
frameworks:
  - PyTorch
language:
  - en
  - zh
downloads: 326
stars: 0
tags:
  - gguf
---

# Ernie-Image-Flex-V1

> Ernie-Image-Flex-V1 - wikeeyang 在 ModelScope 开源的模型。Ernie-Image-Flex-V1

wikeeyang/Ernie-Image-Flex-V1 是 ModelScope 魔搭社区上的text-to-image-synthesis模型，采用 apache-2.0 许可，基于 PaddlePaddle/ERNIE-Image 构建。

- **Repository**: wikeeyang/Ernie-Image-Flex-V1
- **License**: apache-2.0
- **Tasks**: text-to-image-synthesis
- **Base model**: PaddlePaddle/ERNIE-Image
- **Tags**: gguf
- **Downloads**: 326
- **Stars**: 0
- **Last updated**: 2026-05-12

Source: https://www.modelscope.cn/models/wikeeyang/Ernie-Image-Flex-V1

---

# Ernie-Image-Flex V1.0

本模型为百度 [ERNIE-Image](https://modelscope.cn/models/PaddlePaddle/ERNIE-Image) 的微调模型，相比 [ERNIE-Image-Turbo](https://modelscope.cn/models/PaddlePaddle/ERNIE-Image-Turbo) 在以下几方面有所改善和提升 (The model is a fine-tuned version of Baidu's [ERNIE-Image](https://modelscope.cn/models/PaddlePaddle/ERNIE-Image), with improvements and enhancements over [ERNIE-Image-Turbo](https://modelscope.cn/models/PaddlePaddle/ERNIE-Image-Turbo) in the following aspects)：

1. 基本去除  [ERNIE-Image-Turbo](https://modelscope.cn/models/PaddlePaddle/ERNIE-Image-Turbo) 8 步模型油腻较重、噪点过强、颜色偏黄的问题。(Essentially eliminates the issues of heavy oiliness, excessive noise, and yellowish color found in the 8-step [ERNIE-Image-Turbo](https://modelscope.cn/models/PaddlePaddle/ERNIE-Image-Turbo) model.)
2. 相比  [ERNIE-Image](https://modelscope.cn/models/PaddlePaddle/ERNIE-Image) 需要 cfg=4.0 / steps=50 出图很慢，本模型只需 cfg=1.0 / steps=16 - 25，且可使用 Turbo LoRA 加速至 step=8，有更好的弹性尺度。(Compared to [ERNIE-Image](https://modelscope.cn/models/PaddlePaddle/ERNIE-Image), which requires cfg=4.0 / steps=50 and is very slow to generate images, this model only needs cfg=1.0 / steps=16-25, and can be accelerated to step=8 using Turbo LoRA, offering better scalability.)
3. 更好的人像构图与人脸优化，画面更干净，复杂构图时肢体相比 [ERNIE-Image-Turbo](https://modelscope.cn/models/PaddlePaddle/ERNIE-Image-Turbo) 不容易崩。(Improved portrait composition and face optimization, resulting in cleaner images, and limbs are less likely to deform in complex compositions compared to [ERNIE-Image-Turbo](https://modelscope.cn/models/PaddlePaddle/ERNIE-Image-Turbo).)
4. 中英文文字渲染能力得到较好延续，比官方原版略有损耗。(Continues to render Chinese and English text well, with slightly less accuracy than the official original version.)
5. 量化模型采用全新的分层/分块/混合精度量化机制，使 8bit 模型与 bf16 模型效果更接近。(The quantized model uses a new hierarchical / block / mixed-precision quantization mechanism, making the 8bit model closer in effect to the bf16 model.)

本模型依旧为纯净底模，可用于微调或 LoRA 训练，也可叠加各种 Turbo / Distill / NSFW / SFW 的 LoRA 使用。(Pure base model, suitable for fine-tuning or LoRA training, and can also be combined with various Turbo / Distill / SFW / NSFW LoRAs.)

感谢：**@亮亮rayne**，[@GuangyuanSD](https://huggingface.co/GuangyuanSD)  在本模型调制优化过程中给与技术、构图、美学等方面的指导和支持。

Much thanks to: **@亮亮rayne**, [@GuangyuanSD](https://huggingface.co/GuangyuanSD) for their guidance and support in technical, compositional, and aesthetic aspects during the finetuning of the model.

感谢：[@T8star-Aix](https://space.bilibili.com/385085361)，[@AI-KSK](https://space.bilibili.com/110353151)，[@黑鹤001](https://space.bilibili.com/515231056)，[@大麦茶看AI](https://space.bilibili.com/310599757)，[@joeMultimedia](https://space.bilibili.com/612740284) , [山竹老师AIGC](https://space.bilibili.com/158424637) , [CommintAI](https://space.bilibili.com/158429672) , [雪煮松风](https://space.bilibili.com/692784900) , [冰缘排骨](https://space.bilibili.com/3546582953757144) , [活力网TV](https://space.bilibili.com/92055552) , [HyLan_L](https://space.bilibili.com/591877893) , 等一众大佬对本人微调各种模型的技术评测、使用介绍、社区推荐及进一步微调提升等方面的建议和支持。

Much thanks to: [@T8star-Aix](https://space.bilibili.com/385085361), [@AI-KSK](https://space.bilibili.com/110353151), [@黑鹤001](https://space.bilibili.com/515231056), [@大麦茶看AI](https://space.bilibili.com/310599757), [@joeMultimedia](https://space.bilibili.com/612740284), [山竹老师AIGC](https://space.bilibili.com/158424637), [CommintAI](https://space.bilibili.com/158429672), [雪煮松风](https://space.bilibili.com/692784900), [冰缘排骨](https://space.bilibili.com/3546582953757144), [活力网TV](https://space.bilibili.com/92055552), [HyLan_L](https://space.bilibili.com/591877893), and many other experts for their advice and support in technical evaluation, usage introduction, community recommendations, and further fine-tuning improvements of various models I worked on.

<p align="center">
    <img src="v1-sample.jpg" width="1200"/>
<p>



基础工作流请参见(Base Workflow)：[v1-example-workflow.json](https://www.modelscope.cn/models/wikeeyang/Ernie-Image-Flex-V1/resolve/master/v1-example-workflow.json)

Also on: [Huggingface](https://huggingface.co/wikeeyang/Ernie-Image-Flex-V1)

The model under Apache 2.0.