tomesd
Token merger
An optimization technique applied to Stable Diffusion images by merging redundant tokens to reduce computational requirements
Speed up Stable Diffusion with this one simple trick!
1k stars
20 watching
80 forks
Language: Python
last commit: almost 3 years agoLinked from 1 awesome list
amrzv/awesome-colab-notebooksRelated projects:
| Repository | Description | Stars |
|---|---|---|
 yxuansu/tacl | Improves pre-trained language models by encouraging an isotropic and discriminative distribution of token representations. | 92 |
 shellward/stablenderdiffusion | Enables seamless integration of Blender 3D creation and Stable Diffusion AI image generation tools. | 75 |
 diffusionkinetics/open | A monorepo of Haskell projects focused on data science and machine learning | 116 |
 xujiajun/gotokenizer | A tokenizer based on dictionary and Bigram language models for text segmentation in Chinese | 21 |
 madebyollin/taesd | A Tiny AutoEncoder implementation for Stable Diffusion's image generation model | 602 |
 albarji/mixture-of-diffusers | Generates detailed images by combining multiple diffusion processes focused on different regions of the image. | 418 |
 springdaisy/gbdt | An implementation of Gradient Boosted Decision Trees with sparse output for high-dimensional data | 0 |
 jquesnelle/txt2imghd | Creates detailed images by upscaling and refining Stable Diffusion outputs | 695 |
 blueturtleai/gimp-stable-diffusion | An open-source plugin for GIMP image editing software to generate images using stable-diffusion models. | 331 |
 sddebz/stable-diffusion-krita-plugin | An interface for Krita's Stable Diffusion plugin | 423 |
 c4n/pythonlexto | A Python wrapper around the Thai word segmentator LexTo, allowing developers to easily integrate it into their applications. | 1 |
 maks-s/sd-akashic | A compendium of resources and tools for exploring the capabilities and limitations of Stable Diffusion | 1,638 |
 yfzhang114/llava-align | Debiasing techniques to minimize hallucinations in large visual language models | 75 |
 davisyoshida/lorax | A JAX transform that simplifies the training of large language models by reducing memory usage through low-rank adaptation. | 134 |
 lingyanruan/lakdnet | An image deblurring model that revisits the use of convolutional neural networks to improve efficiency and performance compared to transformers. | 102 |