ContainersNVIDIADeepSeek-R1NVIDIALlama-3.1-Nemotron-70B-InstructNVIDIAPyTorch
CollectionsNVIDIAOmniverse Kit (FB)NVIDIADeepStream SDKNVIDIAOmniverse Kit App Streaming
ModelsNVIDIAStyleGAN3 pretrained modelsNVIDIAPeopleNetNVIDIATrafficCamNet
ResourcesNVIDIARiva Skills Quick Start
Helm ChartsNVIDIAGPU OperatorNVIDIANVIDIA NIM Operator
NVIDIA Deep Learning Examples
NVIDIA Deep Learning Examples
ResNet50 v1.5 for PaddlePaddle
Resource
NVIDIA Deep Learning Examples
NVIDIA Deep Learning Examples
ResNet50 v1.5 for PaddlePaddle

With modified architecture and initialization this ResNet50 version gives ~0.5% better accuracy than original.

Changelog

  1. December 2021
  • Initial release
  • Cosine LR schedule
  • DALI support
  • DALI-CPU dataloader
  • Added A100 scripts
  • Paddle AMP
  1. January 2022
  • Added options Label Smoothing, fan-in initialization, skipping weight decay on batch norm gamma and bias.
  • Updated README
  • A100 convergence benchmark

Known issues

  • Allreduce issues to top1 and top5 accuracy in evaluation. Workaround: use build_strategy.fix_op_run_order = True for eval program. (refer to Paddle-issue-39567 for details)