search

HabanaAI / Model-References

Reference models for Intel(R) Gaudi(R) AI Accelerator
155 stars 81 forks source link

readme

Intel® Gaudi® AI Accelerator Examples for Training and Inference

Model List and Performance Data

Please visit this page for performance information.

This repository is a collection of models that have been ported to run on Intel Gaudi AI accelerator. They are intended as examples, and will be reasonably optimized for performance while still being easy to read.

Computer Vision

Models Framework Validated on Gaudi Validated on Gaudi 2
ResNet50, ResNeXt101 PyTorch Training Training, Inference
ResNet152 PyTorch Training -
MobileNetV2 PyTorch Training -
UNet 2D, Unet3D PyTorch Lightning Training, Inference Training, Inference
SSD PyTorch Training Training
GoogLeNet PyTorch Training -
Vision Transformer PyTorch Training -
DINO PyTorch Training -
YOLOX PyTorch Training -

Natural Language Processing

Models Framework Validated on Gaudi Validated on Gaudi 2
BERT Pretraining and Finetuning PyTorch Training, Inference Training, Inference
DeepSpeed BERT-1.5B, BERT-5B PyTorch Training -
BART PyTorch Training -

Audio

Models Framework Validated on Gaudi Validated on Gaudi 2
Wav2Vec2ForCTC PyTorch Inference Inference

Generative Models

Models Framework Validated on Gaudi Validated on Gaudi 2
Stable Diffusion PyTorch Lightning Training Training
Stable Diffusion FineTuning PyTorch Training Training

MLPerf™ Training 4.0

Models Framework Validated on Gaudi Validated on Gaudi 2
GPT3 PyTorch - Training
Llama 70B LoRA PyTorch - Training

MLPerf™ Inference 4.0

Models Framework Validated on Gaudi Validated on Gaudi 2
Llama 70B PyTorch - Inference
Stable Diffusion XL PyTorch - Inference

MLPerf™ is a trademark and service mark of MLCommons Association in the United States and other countries. All rights reserved. Unauthorized use is strictly prohibited.

Reporting Bugs/Feature Requests

We welcome you to use the GitHub issue tracker to report bugs or suggest features.

When filing an issue, please check existing open, or recently closed, issues to make sure somebody else hasn't already reported the issue. Please try to include as much information as you can. Details like these are incredibly useful:

Community

Hugging Face

Megatron-DeepSpeed

DeepSpeed-Chat

Fairseq