|
--- |
|
license: apache-2.0 |
|
license_name: kohaku-license-1.0 |
|
datasets: |
|
- laion/conceptual-captions-12m-webdataset |
|
- CaptionEmporium/coyo-hd-11m-llavanext |
|
- KBlueLeaf/danbooru2023-metadata-database |
|
- graph-based-captions/GBC10M |
|
language: |
|
- en |
|
pipeline_tag: text-generation |
|
library_name: transformers |
|
--- |
|
# TIPO: Text to Image with text presampling for Prompt Optimization |
|
|
|
***This 100M model is still under development*** |
|
|
|
100M LLaMA arch model trained for TIPO.<br> |
|
Tech Report: https://arxiv.org/abs/2411.08127 |
|
|
|
 |
|
|
|
## Introduction |
|
|
|
In this project, we introduce "TIPO" (**T**ext to **I**mage with text presampling for **P**rompt **O**ptimization), an innovative framework designed to significantly enhance the quality and usability of Text-to-Image (T2I) generative models. TIPO utilizes the Large Language Models (LLMs) to perform "Text Presampling" within the inference pipeline of text-to-image generative modeling. By refining and extending user input prompts, TIPO enables generative models to produce superior results with minimal user effort, making T2I systems more accessible and effective for a wider range of users. |
|
|
|
## Usage |
|
Use updated version of DTG extension (renamed to z-tipo-extension), current version of z-tipo-extension support stable-diffusion-webui, stable-diffusion-webui-forge and ComfyUI. SD-Next haven't been tested. |
|
https://github.com/KohakuBlueleaf/z-tipo-extension |
|
|
|
## Model arch and Training |
|
|
|
This model is LLaMA arch with 200M parameters, the training data is combined version of Danbooru2023, Coyo-HD-11M. <br> |
|
The total token seen is around 50B tokens. <br> |
|
For more information please refer to the tech report and following table. |
|
|
|
| | TIPO-200M | TIPO-200M-ft | TIPO-500M | |
|
| ----------------- | ------------------------------------------------------------------------------ | ---------------------------------- | ------------------------------------------------------------------------------ | |
|
| Arch | LLaMA | LLaMA | LLaMA | |
|
| Max ctx length | 1024 | 1024 | 1024 | |
|
| Batch Size | 2048 | 2048 | 3584 | |
|
| Training dataset | Danbooru, GBC10M, 5epoch<br />Danbooru, GBC10M, Coyo11M, 3epoch | Danbooru(pixtral), Coyo11M, 2epoch | Danbooru, GBC10M, Coyo11M, 5epoch | |
|
| Real Token Seen* | 40B token | 50B (10B more from TIPO-200M) | 30B token | |
|
| Training Hardware | RTX 3090 x 4 | RTX 3090 x 4 | H100 x 8 | |
|
| Training Time | 420 hour` | 120 hour` | 100 hour` | |
|
| Huggingface | [KBlueLeaf/TIPO-200M · Hugging Face](https://huggingface.co/KBlueLeaf/TIPO-200M) | [KBlueLeaf/TIPO-200M-ft · Hugging Face](https://huggingface.co/KBlueLeaf/TIPO-200M-ft) | [KBlueLeaf/TIPO-500M · Hugging Face](https://huggingface.co/KBlueLeaf/TIPO-500M) | |
|
|
|
*: We only count "non-padding token" in the token seen, since all the training data have very large length range. <br> |
|
`: Since the training data is pretty short, it cost more time to reach same token seen than general LLM pretraining. <br> |
|
As reference, with 4096 as max ctx length and almost all the data have reach that length, you may only need 2days to reach 10B token seen on RTX 3090 x 4 with 200M model. |
|
|
|
### Evaluation |
|
**Evaluation are done on TIPO-200M model** <br> |
|
We have tested TIPO compared to other Model in several test and metrics: |
|
|
|
#### Scenery tag test |
|
|
|
In this test we use single "scenery" tag as input. (With some certain meta) <br> |
|
To test each prompt gen method to see if they can obtain the desired distribution of outputs while maintain the quality of images. |
|
|
|
| Scenery Tag Test | Original | GPT4o-mini | Prompt DB | Promptis | TIPO(ours) | |
|
| ---- | ---- | ---- | ---- | ---- | ---- | |
|
| FDD ↓ | 0.3558 | 0.5414 | 0.3247 | *0.2350* | **0.2282** | |
|
| Aesthetic ↑ | 5.0569 | **6.3676** | 6.1609 | 5.9468 | *6.2571* | |
|
| AI Corrupt ↑ | 0.4257 | *0.7490* | 0.5024 | 0.5669 | **0.9195** | |
|
|
|
#### Short/Truncated Long test |
|
|
|
In this test we use short caption or manually truncated caption from GBC10M and CoyoHD11M. <br> |
|
This test examine the ability of prompt gen method on handling almostly completed prompts. |
|
|
|
| Short | Original | GPT4o-mini | Prompt DB | Promptis | TIPO(ours) | |
|
| ---- | ---- | ---- | ---- | ---- | ---- | |
|
| FDD ↓ | 0.0957 | 0.1668 | *0.0980* | 0.1783 | 0.1168 | |
|
| Aesthetic ↑ | 5.8370 | **6.0589** | 5.8213 | 5.7963 | *5.8531* | |
|
| AI Corrupt ↑ | 0.7113 | 0.6985 | 0.7064 | 0.6314 | **0.7131** | |
|
|
|
| Truncated Long | Original | GPT4o-mini | Prompt DB | Promptis | TIPO(ours) | |
|
| ---- | ---- | ---- | ---- | ---- | ---- | |
|
| FDD ↓ | 0.0955 | 0.1683 | *0.1247* | 0.2096 | 0.1210 | |
|
| Aesthetic ↑ | 5.7497 | **6.0168** | 5.8191 | 5.7759 | *5.8364* | |
|
| AI Corrupt ↑ | 0.6868 | 0.6712 | 0.6741 | 0.5925 | **0.7130** | |
|
|
|
## LICENSE |
|
For research purpose, this model is released under Apache-2.0 License. |
|
|
|
### Citation |
|
```bibtex |
|
@misc{yeh2024tipotextimagetext, |
|
title={TIPO: Text to Image with Text Presampling for Prompt Optimization}, |
|
author={Shih-Ying Yeh and Sang-Hyun Park and Giyeong Oh and Min Song and Youngjae Yu}, |
|
year={2024}, |
|
eprint={2411.08127}, |
|
archivePrefix={arXiv}, |
|
primaryClass={cs.CV}, |
|
url={https://arxiv.org/abs/2411.08127}, |
|
} |
|
``` |
