2 个仓库
Awesome GitHub RepositoriesInference Speed Optimization
Training and distillation methods designed to reduce the temporal cost of generating samples from generative models.
Distinct from Distributed Consistency Models: Distinct from Distributed Consistency Models: focuses on reducing generation latency in AI models rather than data consistency in distributed systems.
Explore 2 awesome GitHub repositories matching software engineering & architecture · Inference Speed Optimization. Refine with filters or upvote what's useful.
Awesome Inference Speed Optimization GitHub Repositories
用 AI 发现最棒的仓库。我们将通过 AI 为您搜索最匹配的仓库。
- openai/consistency_models
openai/consistency_models
6,492在 GitHub 上查看This project is a framework for training and sampling generative models designed to produce high-quality images in few steps. It provides implementations for image generation models that transform random noise into structured visual data through an optimized sampling process. The system specializes in accelerating image generation through consistency distillation and consistency training. It includes tools to transform pre-trained diffusion models into faster versions by distilling knowledge from a teacher model into a student model, as well as methods to train consistency models from scratch
Reduces the temporal cost of image creation by training student models to mimic teacher models.
Python - meituan-longcat/longcat-video
meituan-longcat/LongCat-Video
4,460在 GitHub 上查看LongCat-Video 是一系列专门用于视频合成的模型集合，采用基于大语言模型的架构，用于从文本、图像或现有序列创建高分辨率视频。它包括用于文本生成视频、图像生成视频动画以及创建数字人头像的专用系统。 该项目提供了通过视频延续模型延长现有剪辑长度的特定功能，该模型可以预测后续帧。它还支持将角色口型与音频和文本提示同步，以制作说话视频。 该系统结合了多种优化技术来管理生成效率，包括基于蒸馏的采样和量化，以减少内存使用和推理延迟。其他结构组件涵盖了潜在空间压缩和时空建模，以保持跨时间和空间的连贯性。
Reduces generation latency and memory usage through distillation and quantization.
Python
