Update README.md
Browse files
README.md
CHANGED
|
@@ -1,212 +1,212 @@
|
|
| 1 |
-
---
|
| 2 |
-
pipeline_tag: text-to-image
|
| 3 |
-
tags:
|
| 4 |
-
- text-to-image
|
| 5 |
-
- image-generation
|
| 6 |
-
- diffusion-transformer
|
| 7 |
-
- dit
|
| 8 |
-
- comfyui
|
| 9 |
-
- checkpoint
|
| 10 |
-
- safetensors
|
| 11 |
-
- aio
|
| 12 |
-
- all-in-one
|
| 13 |
-
- ernie-image
|
| 14 |
-
- ernie-image-turbo
|
| 15 |
-
- bf16
|
| 16 |
-
- fp8
|
| 17 |
-
license: apache-2.0
|
| 18 |
-
widget:
|
| 19 |
-
- text: "A premium sci-fi poster with readable text, cinematic lighting, and a futuristic city skyline at night."
|
| 20 |
-
- text: "A clean product advertisement for a coffee brand with readable package text, elegant layout, and warm morning light."
|
| 21 |
-
---
|
| 22 |
-
|
| 23 |
-
# ERNIE AIO for ComfyUI
|
| 24 |
-
**All-in-One Checkpoints • Base + Turbo • BF16 & FP8 • Integrated Text Encoder + VAE**
|
| 25 |
-
|
| 26 |
-
This repository provides **AIO (All-in-One)** checkpoint versions of **ERNIE-Image** and **ERNIE-Image-Turbo** for **ComfyUI**.
|
| 27 |
-
|
| 28 |
-
A ready-to-use **ComfyUI workflow** is included in the **`Workflow/`** folder, and the optional **Prompt Enhancer** files are included in the **`prompt-enhancer/`** folder.
|
| 29 |
-
|
| 30 |
-
These AIO releases are designed for users who prefer a classic **single-checkpoint workflow**.
|
| 31 |
-
Each checkpoint already contains the **image model**, **text encoder**, and **VAE**, so no separate text encoder or VAE loading is required.
|
| 32 |
-
|
| 33 |
-
### ERNIE-AIO-Base
|
| 34 |
-

|
| 35 |
-
|
| 36 |
-
### ERNIE-AIO-Turbo
|
| 37 |
-

|
| 38 |
-
|
| 39 |
-
## Overview
|
| 40 |
-
This repository provides **AIO (All-in-One)** checkpoint versions of **ERNIE-Image** and **ERNIE-Image-Turbo** for **ComfyUI**.
|
| 41 |
-
|
| 42 |
-
These AIO releases are designed for users who prefer a classic **single-checkpoint workflow**.
|
| 43 |
-
Each checkpoint already contains the **image model**, **text encoder**, and **VAE**, so no separate text encoder or VAE loading is required.
|
| 44 |
-
|
| 45 |
-
|
| 46 |
-
## Available Files
|
| 47 |
-
|
| 48 |
-
### Main checkpoints
|
| 49 |
-
- `ERNIE-AIO-Base-bf16.safetensors`
|
| 50 |
-
- `ERNIE-AIO-Base-fp8.safetensors`
|
| 51 |
-
- `ERNIE-AIO-Turbo-bf16.safetensors`
|
| 52 |
-
- `ERNIE-AIO-Turbo-fp8.safetensors`
|
| 53 |
-
|
| 54 |
-
### Workflow
|
| 55 |
-
- `Workflow/ERNIE-AIO-Upscale.json`
|
| 56 |
-
|
| 57 |
-
### Prompt Enhancer
|
| 58 |
-
- `prompt-enhancer/` contains the optional ERNIE Prompt Enhancer files used by the workflow
|
| 59 |
-
|
| 60 |
-
## What “AIO” means here
|
| 61 |
-
The AIO checkpoints are packaged so they can be placed directly in:
|
| 62 |
-
|
| 63 |
-
`ComfyUI/models/checkpoints`
|
| 64 |
-
|
| 65 |
-
and loaded with the normal **Checkpoint Loader** node in ComfyUI.
|
| 66 |
-
|
| 67 |
-
This means you can use a very small workflow such as:
|
| 68 |
-
|
| 69 |
-
1. **Checkpoint Loader**
|
| 70 |
-
2. **CLIP Text Encode (Prompt)**
|
| 71 |
-
3. **Empty Latent Image**
|
| 72 |
-
4. **KSampler**
|
| 73 |
-
5. **VAE Decode**
|
| 74 |
-
6. **Save Image**
|
| 75 |
-
|
| 76 |
-
A workflow file may also be included separately for convenience.
|
| 77 |
-
|
| 78 |
-
## Model Background
|
| 79 |
-
These AIO releases are based on Baidu’s official **ERNIE-Image** family.
|
| 80 |
-
|
| 81 |
-
According to the official model pages:
|
| 82 |
-
|
| 83 |
-
- **ERNIE-Image** is an open text-to-image model built on a **single-stream Diffusion Transformer (DiT)** and paired with a lightweight **Prompt Enhancer**
|
| 84 |
-
- **ERNIE-Image-Turbo** is the **distilled** variant, designed for strong fidelity in only **8 inference steps**
|
| 85 |
-
- The official releases emphasize **instruction following**, **text rendering**, and **structured image generation**, making them especially interesting for posters, comics, infographics, UI-like layouts, and other text-heavy or layout-sensitive images
|
| 86 |
-
|
| 87 |
-
## AIO Variants
|
| 88 |
-
|
| 89 |
-
### ERNIE-AIO-Base
|
| 90 |
-
The Base version is intended for more deliberate, stable, and higher-quality generation.
|
| 91 |
-
|
| 92 |
-
**Recommended settings:**
|
| 93 |
-
- **Steps:** 50
|
| 94 |
-
- **CFG:** 4.0
|
| 95 |
-
|
| 96 |
-
Use this version when you want stronger refinement, more careful rendering, and the full non-turbo experience.
|
| 97 |
-
|
| 98 |
-
### ERNIE-AIO-Turbo
|
| 99 |
-
The Turbo version is intended for fast generation and quick iteration.
|
| 100 |
-
|
| 101 |
-
**Recommended settings:**
|
| 102 |
-
- **Steps:** 8
|
| 103 |
-
- **CFG:** 1.0
|
| 104 |
-
|
| 105 |
-
Use this version when you want faster turnaround while still keeping good prompt fidelity and strong visual output.
|
| 106 |
-
|
| 107 |
-
## Precision / Quantization Notes
|
| 108 |
-
The **FP8** versions in this repository were quantized by me.
|
| 109 |
-
|
| 110 |
-
For these FP8 AIO checkpoints:
|
| 111 |
-
|
| 112 |
-
- **Image model:** FP8
|
| 113 |
-
- **Text encoder:** FP8
|
| 114 |
-
- **VAE:** FP32
|
| 115 |
-
|
| 116 |
-
This setup was chosen to reduce size and improve practicality while keeping decoding stable.
|
| 117 |
-
|
| 118 |
-
## Prompt Enhancer
|
| 119 |
-
|
| 120 |
-
This repository also includes an optional **Prompt Enhancer** in the `
|
| 121 |
-
|
| 122 |
-
It can be used with the included workflow to automatically expand short prompts into richer and more descriptive prompts.
|
| 123 |
-
This is optional, but useful if you want more detailed prompt wording or a more guided workflow experience.
|
| 124 |
-
|
| 125 |
-
## Prompting
|
| 126 |
-
These models work well with **natural language prompts**.
|
| 127 |
-
|
| 128 |
-
They are especially suitable for:
|
| 129 |
-
- posters
|
| 130 |
-
- title cards
|
| 131 |
-
- product ads
|
| 132 |
-
- readable text in images
|
| 133 |
-
- structured layouts
|
| 134 |
-
- comics
|
| 135 |
-
- infographics
|
| 136 |
-
- UI-like visuals
|
| 137 |
-
- realistic and stylized imagery
|
| 138 |
-
|
| 139 |
-
For best results, describe:
|
| 140 |
-
- subject
|
| 141 |
-
- environment
|
| 142 |
-
- composition
|
| 143 |
-
- mood
|
| 144 |
-
- lighting
|
| 145 |
-
- camera framing
|
| 146 |
-
- text that should appear in the image
|
| 147 |
-
|
| 148 |
-
## Example Prompt Ideas
|
| 149 |
-
- A cinematic movie poster with a bold readable title at the top, dramatic lighting, and a futuristic skyline.
|
| 150 |
-
- A modern coffee advertisement with readable packaging text and a premium warm studio look.
|
| 151 |
-
- A clean infographic with clearly labeled sections and a structured layout.
|
| 152 |
-
- A comic-style scene with readable speech bubbles and strong composition.
|
| 153 |
-
|
| 154 |
-
## ComfyUI Usage
|
| 155 |
-
Place the checkpoints in:
|
| 156 |
-
|
| 157 |
-
`ComfyUI/models/checkpoints`
|
| 158 |
-
|
| 159 |
-
Then load them through **Checkpoint Loader** like a normal checkpoint.
|
| 160 |
-
|
| 161 |
-
No separate:
|
| 162 |
-
- VAE loader
|
| 163 |
-
- text encoder loader
|
| 164 |
-
- extra model assembly
|
| 165 |
-
|
| 166 |
-
is required for the AIO versions.
|
| 167 |
-
|
| 168 |
-
## Repository Structure
|
| 169 |
-
|
| 170 |
-
- Main AIO checkpoints for **ERNIE-AIO-Base** and **ERNIE-AIO-Turbo**
|
| 171 |
-
- `Workflow/` contains the ready-to-use ComfyUI workflow JSON
|
| 172 |
-
- `prompt-enhancer/` contains the optional Prompt Enhancer files used by the workflow
|
| 173 |
-
- `images/` contains the preview / title images for the Base and Turbo releases
|
| 174 |
-
|
| 175 |
-
## Civitai Releases
|
| 176 |
-
|
| 177 |
-
If you prefer the Civitai pages, you can also find the releases here:
|
| 178 |
-
|
| 179 |
-
- Turbo: https://civitai.red/models/2551548?modelVersionId=2868512
|
| 180 |
-
- Base: https://civitai.red/models/2551548?modelVersionId=2867462
|
| 181 |
-
|
| 182 |
-
## Limitations
|
| 183 |
-
As with other image generation models:
|
| 184 |
-
|
| 185 |
-
- text rendering can still vary depending on prompt complexity
|
| 186 |
-
- highly dense typography may require retries
|
| 187 |
-
- prompt phrasing can strongly affect layout quality
|
| 188 |
-
- FP8 variants may behave slightly differently from BF16 variants depending on hardware and workflow
|
| 189 |
-
|
| 190 |
-
## Intended Use
|
| 191 |
-
These releases are intended for:
|
| 192 |
-
- local image generation in ComfyUI
|
| 193 |
-
- poster and design-style generation
|
| 194 |
-
- readable title images
|
| 195 |
-
- structured compositions
|
| 196 |
-
- users who want a simple all-in-one checkpoint workflow
|
| 197 |
-
|
| 198 |
-
## Credits
|
| 199 |
-
Huge thanks to the original **ERNIE-Image** team at **Baidu** for releasing the official models.
|
| 200 |
-
|
| 201 |
-
Official sources:
|
| 202 |
-
- **ERNIE-Image:** https://huggingface.co/baidu/ERNIE-Image
|
| 203 |
-
- **ERNIE-Image-Turbo:** https://huggingface.co/baidu/ERNIE-Image-Turbo
|
| 204 |
-
- **GitHub:** https://github.com/baidu/ERNIE-Image
|
| 205 |
-
|
| 206 |
-
## License
|
| 207 |
-
This repository is based on the official ERNIE-Image releases, whose Hugging Face model pages list the license as **Apache-2.0**.
|
| 208 |
-
|
| 209 |
-
## Final Note
|
| 210 |
-
These AIO releases were made to simplify the ERNIE workflow for ComfyUI users.
|
| 211 |
-
|
| 212 |
Load checkpoint, write prompt, sample, decode, save.
|
|
|
|
| 1 |
+
---
|
| 2 |
+
pipeline_tag: text-to-image
|
| 3 |
+
tags:
|
| 4 |
+
- text-to-image
|
| 5 |
+
- image-generation
|
| 6 |
+
- diffusion-transformer
|
| 7 |
+
- dit
|
| 8 |
+
- comfyui
|
| 9 |
+
- checkpoint
|
| 10 |
+
- safetensors
|
| 11 |
+
- aio
|
| 12 |
+
- all-in-one
|
| 13 |
+
- ernie-image
|
| 14 |
+
- ernie-image-turbo
|
| 15 |
+
- bf16
|
| 16 |
+
- fp8
|
| 17 |
+
license: apache-2.0
|
| 18 |
+
widget:
|
| 19 |
+
- text: "A premium sci-fi poster with readable text, cinematic lighting, and a futuristic city skyline at night."
|
| 20 |
+
- text: "A clean product advertisement for a coffee brand with readable package text, elegant layout, and warm morning light."
|
| 21 |
+
---
|
| 22 |
+
|
| 23 |
+
# ERNIE AIO for ComfyUI
|
| 24 |
+
**All-in-One Checkpoints • Base + Turbo • BF16 & FP8 • Integrated Text Encoder + VAE**
|
| 25 |
+
|
| 26 |
+
This repository provides **AIO (All-in-One)** checkpoint versions of **ERNIE-Image** and **ERNIE-Image-Turbo** for **ComfyUI**.
|
| 27 |
+
|
| 28 |
+
A ready-to-use **ComfyUI workflow** is included in the **`Workflow/`** folder, and the optional **Prompt Enhancer** files are included in the **`prompt-enhancer/`** folder.
|
| 29 |
+
|
| 30 |
+
These AIO releases are designed for users who prefer a classic **single-checkpoint workflow**.
|
| 31 |
+
Each checkpoint already contains the **image model**, **text encoder**, and **VAE**, so no separate text encoder or VAE loading is required.
|
| 32 |
+
|
| 33 |
+
### ERNIE-AIO-Base
|
| 34 |
+

|
| 35 |
+
|
| 36 |
+
### ERNIE-AIO-Turbo
|
| 37 |
+

|
| 38 |
+
|
| 39 |
+
## Overview
|
| 40 |
+
This repository provides **AIO (All-in-One)** checkpoint versions of **ERNIE-Image** and **ERNIE-Image-Turbo** for **ComfyUI**.
|
| 41 |
+
|
| 42 |
+
These AIO releases are designed for users who prefer a classic **single-checkpoint workflow**.
|
| 43 |
+
Each checkpoint already contains the **image model**, **text encoder**, and **VAE**, so no separate text encoder or VAE loading is required.
|
| 44 |
+
|
| 45 |
+
|
| 46 |
+
## Available Files
|
| 47 |
+
|
| 48 |
+
### Main checkpoints
|
| 49 |
+
- `ERNIE-AIO-Base-bf16.safetensors`
|
| 50 |
+
- `ERNIE-AIO-Base-fp8.safetensors`
|
| 51 |
+
- `ERNIE-AIO-Turbo-bf16.safetensors`
|
| 52 |
+
- `ERNIE-AIO-Turbo-fp8.safetensors`
|
| 53 |
+
|
| 54 |
+
### Workflow
|
| 55 |
+
- `Workflow/ERNIE-AIO-Upscale.json`
|
| 56 |
+
|
| 57 |
+
### Prompt Enhancer
|
| 58 |
+
- `prompt-enhancer/` contains the optional ERNIE Prompt Enhancer files used by the workflow
|
| 59 |
+
|
| 60 |
+
## What “AIO” means here
|
| 61 |
+
The AIO checkpoints are packaged so they can be placed directly in:
|
| 62 |
+
|
| 63 |
+
`ComfyUI/models/checkpoints`
|
| 64 |
+
|
| 65 |
+
and loaded with the normal **Checkpoint Loader** node in ComfyUI.
|
| 66 |
+
|
| 67 |
+
This means you can use a very small workflow such as:
|
| 68 |
+
|
| 69 |
+
1. **Checkpoint Loader**
|
| 70 |
+
2. **CLIP Text Encode (Prompt)**
|
| 71 |
+
3. **Empty Latent Image**
|
| 72 |
+
4. **KSampler**
|
| 73 |
+
5. **VAE Decode**
|
| 74 |
+
6. **Save Image**
|
| 75 |
+
|
| 76 |
+
A workflow file may also be included separately for convenience.
|
| 77 |
+
|
| 78 |
+
## Model Background
|
| 79 |
+
These AIO releases are based on Baidu’s official **ERNIE-Image** family.
|
| 80 |
+
|
| 81 |
+
According to the official model pages:
|
| 82 |
+
|
| 83 |
+
- **ERNIE-Image** is an open text-to-image model built on a **single-stream Diffusion Transformer (DiT)** and paired with a lightweight **Prompt Enhancer**
|
| 84 |
+
- **ERNIE-Image-Turbo** is the **distilled** variant, designed for strong fidelity in only **8 inference steps**
|
| 85 |
+
- The official releases emphasize **instruction following**, **text rendering**, and **structured image generation**, making them especially interesting for posters, comics, infographics, UI-like layouts, and other text-heavy or layout-sensitive images
|
| 86 |
+
|
| 87 |
+
## AIO Variants
|
| 88 |
+
|
| 89 |
+
### ERNIE-AIO-Base
|
| 90 |
+
The Base version is intended for more deliberate, stable, and higher-quality generation.
|
| 91 |
+
|
| 92 |
+
**Recommended settings:**
|
| 93 |
+
- **Steps:** 50
|
| 94 |
+
- **CFG:** 4.0
|
| 95 |
+
|
| 96 |
+
Use this version when you want stronger refinement, more careful rendering, and the full non-turbo experience.
|
| 97 |
+
|
| 98 |
+
### ERNIE-AIO-Turbo
|
| 99 |
+
The Turbo version is intended for fast generation and quick iteration.
|
| 100 |
+
|
| 101 |
+
**Recommended settings:**
|
| 102 |
+
- **Steps:** 8
|
| 103 |
+
- **CFG:** 1.0
|
| 104 |
+
|
| 105 |
+
Use this version when you want faster turnaround while still keeping good prompt fidelity and strong visual output.
|
| 106 |
+
|
| 107 |
+
## Precision / Quantization Notes
|
| 108 |
+
The **FP8** versions in this repository were quantized by me.
|
| 109 |
+
|
| 110 |
+
For these FP8 AIO checkpoints:
|
| 111 |
+
|
| 112 |
+
- **Image model:** FP8
|
| 113 |
+
- **Text encoder:** FP8
|
| 114 |
+
- **VAE:** FP32
|
| 115 |
+
|
| 116 |
+
This setup was chosen to reduce size and improve practicality while keeping decoding stable.
|
| 117 |
+
|
| 118 |
+
## Prompt Enhancer
|
| 119 |
+
|
| 120 |
+
This repository also includes an optional **Prompt Enhancer** in the `text encoder/` folder.
|
| 121 |
+
|
| 122 |
+
It can be used with the included workflow to automatically expand short prompts into richer and more descriptive prompts.
|
| 123 |
+
This is optional, but useful if you want more detailed prompt wording or a more guided workflow experience.
|
| 124 |
+
|
| 125 |
+
## Prompting
|
| 126 |
+
These models work well with **natural language prompts**.
|
| 127 |
+
|
| 128 |
+
They are especially suitable for:
|
| 129 |
+
- posters
|
| 130 |
+
- title cards
|
| 131 |
+
- product ads
|
| 132 |
+
- readable text in images
|
| 133 |
+
- structured layouts
|
| 134 |
+
- comics
|
| 135 |
+
- infographics
|
| 136 |
+
- UI-like visuals
|
| 137 |
+
- realistic and stylized imagery
|
| 138 |
+
|
| 139 |
+
For best results, describe:
|
| 140 |
+
- subject
|
| 141 |
+
- environment
|
| 142 |
+
- composition
|
| 143 |
+
- mood
|
| 144 |
+
- lighting
|
| 145 |
+
- camera framing
|
| 146 |
+
- text that should appear in the image
|
| 147 |
+
|
| 148 |
+
## Example Prompt Ideas
|
| 149 |
+
- A cinematic movie poster with a bold readable title at the top, dramatic lighting, and a futuristic skyline.
|
| 150 |
+
- A modern coffee advertisement with readable packaging text and a premium warm studio look.
|
| 151 |
+
- A clean infographic with clearly labeled sections and a structured layout.
|
| 152 |
+
- A comic-style scene with readable speech bubbles and strong composition.
|
| 153 |
+
|
| 154 |
+
## ComfyUI Usage
|
| 155 |
+
Place the checkpoints in:
|
| 156 |
+
|
| 157 |
+
`ComfyUI/models/checkpoints`
|
| 158 |
+
|
| 159 |
+
Then load them through **Checkpoint Loader** like a normal checkpoint.
|
| 160 |
+
|
| 161 |
+
No separate:
|
| 162 |
+
- VAE loader
|
| 163 |
+
- text encoder loader
|
| 164 |
+
- extra model assembly
|
| 165 |
+
|
| 166 |
+
is required for the AIO versions.
|
| 167 |
+
|
| 168 |
+
## Repository Structure
|
| 169 |
+
|
| 170 |
+
- Main AIO checkpoints for **ERNIE-AIO-Base** and **ERNIE-AIO-Turbo**
|
| 171 |
+
- `Workflow/` contains the ready-to-use ComfyUI workflow JSON
|
| 172 |
+
- `prompt-enhancer/` contains the optional Prompt Enhancer files used by the workflow
|
| 173 |
+
- `images/` contains the preview / title images for the Base and Turbo releases
|
| 174 |
+
|
| 175 |
+
## Civitai Releases
|
| 176 |
+
|
| 177 |
+
If you prefer the Civitai pages, you can also find the releases here:
|
| 178 |
+
|
| 179 |
+
- Turbo: https://civitai.red/models/2551548?modelVersionId=2868512
|
| 180 |
+
- Base: https://civitai.red/models/2551548?modelVersionId=2867462
|
| 181 |
+
|
| 182 |
+
## Limitations
|
| 183 |
+
As with other image generation models:
|
| 184 |
+
|
| 185 |
+
- text rendering can still vary depending on prompt complexity
|
| 186 |
+
- highly dense typography may require retries
|
| 187 |
+
- prompt phrasing can strongly affect layout quality
|
| 188 |
+
- FP8 variants may behave slightly differently from BF16 variants depending on hardware and workflow
|
| 189 |
+
|
| 190 |
+
## Intended Use
|
| 191 |
+
These releases are intended for:
|
| 192 |
+
- local image generation in ComfyUI
|
| 193 |
+
- poster and design-style generation
|
| 194 |
+
- readable title images
|
| 195 |
+
- structured compositions
|
| 196 |
+
- users who want a simple all-in-one checkpoint workflow
|
| 197 |
+
|
| 198 |
+
## Credits
|
| 199 |
+
Huge thanks to the original **ERNIE-Image** team at **Baidu** for releasing the official models.
|
| 200 |
+
|
| 201 |
+
Official sources:
|
| 202 |
+
- **ERNIE-Image:** https://huggingface.co/baidu/ERNIE-Image
|
| 203 |
+
- **ERNIE-Image-Turbo:** https://huggingface.co/baidu/ERNIE-Image-Turbo
|
| 204 |
+
- **GitHub:** https://github.com/baidu/ERNIE-Image
|
| 205 |
+
|
| 206 |
+
## License
|
| 207 |
+
This repository is based on the official ERNIE-Image releases, whose Hugging Face model pages list the license as **Apache-2.0**.
|
| 208 |
+
|
| 209 |
+
## Final Note
|
| 210 |
+
These AIO releases were made to simplify the ERNIE workflow for ComfyUI users.
|
| 211 |
+
|
| 212 |
Load checkpoint, write prompt, sample, decode, save.
|