deepspeedai/DeepSpeedExamples
Example models using DeepSpeed
What it solves
This repository provides a centralized collection of practical implementations and benchmarks to help users leverage the DeepSpeed library for large-scale AI model development.
How it works
It organizes a variety of reference implementations across five key areas: training and fine-tuning, high-performance inference (via DeepSpeed-MII and FastGen), model compression, end-to-end applications, and performance benchmarks.
Who it’s for
Developers and researchers who are using DeepSpeed to train, optimize, or deploy large models and need concrete examples of how to implement these features.
Highlights
- Comprehensive examples for both training and fine-tuning.
- Dedicated inference paths for DeepSpeed-MII, FastGen, and Hugging Face integrations.
- Resources for model compression and performance benchmarking.
Related
- Project
deepspeedai/DeepSpeedDeepSpeed is a PyTorch‑based open‑source library that provides memory‑efficient optimizers, multi‑dimensional parallelism, and hardware‑offload engines to train extremely large deep‑learning models (up to hundreds of billions of parameters) on GPUs, CPUs, and specialized accelerators.
- Project
openvinotoolkit/open_model_zooA collection of optimized pre-trained deep learning models and tools for accelerating the development and deployment of high-performance inference applications.
- Project
CerebriumAI/examplesA collection of ready-to-use deployment examples for the Cerebrium platform, covering LLMs, voice agents, image generation, and inference optimization.
- Project
bghira/SimpleTunerA comprehensive training framework for fine-tuning image, video, and audio generative models, featuring a user-friendly web UI and enterprise-grade orchestration.
- Project
jolibrain/deepdetectA deep learning runtime and REST server that provides a unified CLI and API for training and serving models across multiple data modalities.