DiffusionGemma
DiffusionGemma is an experimental open model developed by Google that utilizes text diffusion techniques to generate entire blocks of text simultaneously, achieving up to four times faster text generation on dedicated GPUs compared to traditional autoregressive models. Released under an Apache 2.0 license, this 26-billion parameter Mixture of Experts (MoE) model is designed for researchers and developers working on speed-critical, interactive local workflows such as in-line editing, rapid iteration, and generating non-linear text structures.
on the blogGoogle Releases DiffusionGemma: 4x Faster Text Generation via Diffusion 4 min
DiffusionGemma is an experimental open model developed by Google that utilizes text diffusion techniques to generate entire blocks of text simultaneously, achieving up to four times faster text generation on dedicated GPUs compared to traditional autoregressive models. Released under an Apache 2.0 license, this 26-billion parameter Mixture of Experts (MoE) model is designed for researchers and developers working on speed-critical, interactive local workflows such as in-line editing, rapid iteration, and generating non-linear text structures.
- 01Up to 4x faster text generation on dedicated GPUs
- 02Generates entire blocks of text simultaneously
- 03Bi-directional attention allowing every token to attend to all others
- 04Intelligent self-correction for real-time output refinement
- 05Operates within 18GB VRAM limits of high-end consumer GPUs when quantized
- 06Released under an Apache 2.0 license