|
--- |
|
library_name: diffusers |
|
--- |
|
|
|
# SPRIGHT-T2I Model Card |
|
|
|
The SPRIGHT-T2I model is a text-to-image diffusion model with high spatial coherency. It was first introduced in [Getting it Right: Improving Spatial Consistency in Text-to-Image Models](https://), authored by Agneet Chatterjee, Gabriela Ben Melech Stan, Estelle Aflalo, |
|
Sayak Paul, Dhruba Ghosh, Tejas Gokhale, Ludwig Schmidt, Hannaneh Hajishirzi, Vasudev Lal, Chitta Baral, and Yezhou Yang. |
|
|
|
SPRIGHT-T2I model was finetuned from stable diffusion v2.1 on a subset of the [SPRIGHT dataset](https://huggingface.co/datasets/SPRIGHT-T2I/spright), which contains images and spatially focused captions. Leveraging SPRIGHT, along with efficient training techniques, we achieve state-of-the art performance in generating spatially accurate images from text. |
|
|
|
The training code and more details available in [SPRIGHT-T2I GitHub Repository](https://github.com/orgs/SPRIGHT-T2I). |
|
|
|
A demo is available on [Spaces](https://huggingface.co/spaces/SPRIGHT-T2I/SPRIGHT-T2I). |
|
|
|
Use SPRIGHT-T2I with 🧨 [`diffusers`](https://huggingface.co/SPRIGHT-T2I/spright-t2i-sd2#usage). |
|
|
|
|
|
## Model Details |
|
|
|
- **Developed by:** Agneet Chatterjee, Gabriela Ben Melech Stan, Estelle Aflalo, Sayak Paul, Dhruba Ghosh, Tejas Gokhale, Ludwig Schmidt, Hannaneh Hajishirzi, Vasudev Lal, Chitta Baral, and Yezhou Yang |
|
- **Model type:** Diffusion-based text-to-image generation model with spatial coherency |
|
- **Language(s) (NLP):** English |
|
- **License:** [More Information Needed] |
|
- **Finetuned from model:** [Stable Diffusion v2-1](https://huggingface.co/stabilityai/stable-diffusion-2-1) |
|
|
|
## Usage |
|
|
|
Use the code below to run SPRIGHT-T2I seamlessly and effectively on [🤗's Diffusers library](https://github.com/huggingface/diffusers) . |
|
|
|
```bash |
|
pip install diffusers transformers accelerate scipy safetensors |
|
``` |
|
|
|
Running the pipeline: |
|
|
|
```python |
|
from diffusers import DiffusionPipeline |
|
|
|
pipe_id = "SPRIGHT-T2I/spright-t2i-sd2" |
|
|
|
pipe = DiffusionPipeline.from_pretrained( |
|
pipe_id, |
|
torch_dtype=torch.float16, |
|
use_safetensors=True, |
|
).to("cuda") |
|
|
|
prompt = "a cute kitten is sitting in a dish on a table" |
|
image = pipe(prompt).images[0] |
|
|
|
image.save("kitten_sittin_in_a_dish.png") |
|
``` |
|
|
|
<img src="kitten_sitting_in_a_dish.png" width="300" alt="img"> |
|
|
|
Additional examples that emphasize spatial coherence: |
|
<img src="result_images/visor.png" width="1000" alt="img"> |
|
|
|
## Bias and Limitations |
|
|
|
The biases and limitation as specified in [Stable Diffusion v2-1](https://huggingface.co/stabilityai/stable-diffusion-2-1) apply here as well. |
|
|
|
## Training |
|
|
|
#### Training Data |
|
|
|
Our training and validation set are a subset of the [SPRIGHT dataset](https://huggingface.co/datasets/SPRIGHT-T2I/spright), and consists of 444 and |
|
50 images respectively, randomly sampled in a 50:50 split between LAION-Aesthetics and Segment Anything. Each image is paired with both, a general and a spatial caption |
|
(from SPRIGHT). During fine-tuning, for each image, we randomly choose one of the given caption types in a 50:50 ratio. |
|
|
|
We find that SPRIGHT largely improves upon existing datasets in capturing spatial relationships. |
|
Additionally, we find that training on images containing a large number of objects results in substantial improvements in spatial consistency. |
|
To construct our dataset, we focused on images with object counts larger than 18, utilizing the open-world image tagging model |
|
[Recognize Anything](https://huggingface.co/xinyu1205/recognize-anything-plus-model) to achieve this constraint. |
|
|
|
#### Training Procedure |
|
|
|
Our base model is Stable Diffusion v2.1. We fine-tune the U-Net and the OpenCLIP-ViT/H text-encoder as part of our training for 10,000 steps, with different learning rates. |
|
|
|
- **Training regime:** fp16 mixed precision |
|
- **Optimizer:** AdamW |
|
- **Gradient Accumulations**: 1 |
|
- **Batch:** 4 x 8 = 32 |
|
- **UNet learning rate:** 0.00005 |
|
- **CLIP text-encoder learning rate:** 0.000001 |
|
- **Hardware:** Training was performed using NVIDIA RTX A6000 GPUs and Intel®Gaudi®2 AI accelerators. |
|
|
|
|
|
## Evaluation |
|
|
|
We find that compared to the baseline model SD 2.1, we largely improve the spatial accuracy, while also enhancing the non-spatial aspects associated with a text-to-image model. |
|
|
|
The following table compares our SPRIGHT-T2I model with SD 2.1 across multiple spatial reasoning and image quality: |
|
|
|
|Method |OA(%) ↑|VISOR-4(%) ↑|T2I-CompBench ↑|FID ↓|CCMD ↓| |
|
|------------------|-------|------------|---------------|-----|------| |
|
|SD v2.1 |47.83 |4.70 |0.1507 |27.39|1.060 | |
|
|SPRIGHT-T2I (ours)|60.68 |16.15 |0.2133 |27.82|0.512 | |
|
|
|
Our key findings are: |
|
- Increased the Object Accuracy (OA) score by 26.86%, indicating that we are much better at generating objects mentioned in the input prompt |
|
- Visor-4 score of 16.15% denotes that for a given input prompt, we consistently generate a spatially accurate image |
|
- Improve on all aspects of the VISOR score while improving the ZS-FID and CMMD score on COCO-30K images by 23.74% and 51.69%, respectively |
|
- Enhance the ability to generate 1 and 2 objects, along with generating the correct number of objects, as indicated by evaluation on the [GenEval](https://github.com/djghosh13/geneval) benchmark. |
|
|
|
### Model Sources |
|
|
|
- **Repository:** [SPRIGHT-T2I GitHub Repository](https://github.com/orgs/SPRIGHT-T2I) |
|
- **Paper:** [Getting it Right: Improving Spatial Consistency in Text-to-Image Models](https://) |
|
- **Demo:** [SPRIGHT-T2I on Spaces](https://huggingface.co/spaces/SPRIGHT-T2I/SPRIGHT-T2I) |
|
|
|
## Citation |
|
|
|
Coming soon |
|
|
|
|
|
|
|
|
|
|
