引入双子座扩散

我们最先进的实验文本扩散模型

output: Gemini Diffusion is an experimental text diffusion model from Google AI, combining the strengths of large-language models with a new generation technique called diffusion.

输出：Gemini扩散是Google AI的实验文本扩散模型，将大语模型的强度与称为扩散的新一代技术相结合。

Traditional language models generate text one token at a time using autoregressive methods. However, in tasks like editing, inserting content in the context of math or code, and merging different ideas, a more flexible and powerful generation paradigm is needed.

传统语言模型一次使用自回归方法一次产生文本一个令牌。但是，在诸如编辑，在数学或代码的上下文中插入内容之类的任务以及合并不同的想法，需要一个更灵活，更强大的生成范式。

Diffusion models offer an alternative approach. Instead of learning to predict text directly, they excel at refining noise to generate outputs step-by-step. This property is leveraged in Gemini Diffusion for efficient generation.

扩散模型提供了另一种方法。他们没有学习直接预测文本，而是在精炼噪声方面表现出色，以逐步产生输出。该特性在双子座扩散中杠杆作用，以进行有效的产生。

Capabilities

功能

Rapid response

快速响应

Generates content significantly faster than even our fastest model so far.

到目前为止，生成内容的速度明显快。

More coherent text

更连贯的文字

Generates entire blocks of tokens at once, meaning it responds more coherently to a user’s prompt than autoregressive models.

一次生成整个令牌块，这意味着它对用户提示的响应比自回归模型更连贯。

Iterative refinement

迭代精致

Corrects errors during generation for more consistent outputs.

在生成过程中纠正错误以获得更一致的输出。

Benchmarks

基准

Gemini Diffusion’s external benchmark performance is comparable to much larger models, whilst also being faster.

双子扩散的外部基准性能与更大的模型相当，同时也更快。

Methodology

方法论

All scores are pass @1 (no majority voting). The Gemini 2.0 Flash-Lite experiments are run with the AI Studio API for the model-id gemini-2.0-flash-lite with the default sampling settings.

所有分数都是通过 @1（没有多数投票）。 GEMINI 2.0闪存 - 闪存实验是使用默认采样设置的Model-ID GEMINI-2.0-FLASH-LITE的AI Studio API运行的。

* Non-agentic evaluation (single turn edit only), max prompt length of 32K.

*非主张评估（仅单转编辑），最大提示长度为32K。

Gemini Diffusion speed

双子扩散速度

Average sampling speed across reported evals.

跨报告的vals的平均采样速度。

Try Gemini Diffusion

尝试双子座扩散

Gemini Diffusion is currently available as an experimental demo to help develop and refine future models. If you're interested in getting access to the demo, please sign up to the waitlist.

双子座扩散目前可作为实验演示可用，以帮助开发和完善未来模型。如果您有兴趣访问演示，请注册候补名单。