• Search by Module
• Search by Words
• Search Projects

• Java
• Python
• JavaScript
• TypeScript
• C++
• Scala
• Blog

Project: imagenet-fast (GitHub Link)

• imagenet-fast-master
  • cifar10
    • train_cifar10.py
    • fp16util.py
    • distributed.py
    • models
      • resnext_101_32x4d.py
      • cifar10
        • main_kuangliu.py
        • clr_resnet.py
        • main.sh
        • __init__.py
        • utils.py
        • main_dxy.py
        • utils_kuangliu.py
        • resnext.py
        • preact_resnet.py
        • senet.py
      • dpn.py
      • wideresnet_new.py
      • se_resnet.py
      • resnet.py
      • wideresnet_old.py
      • nasnet.py
      • wideresnet.py
      • resnext_50_32x4d.py
      • convert_torch.py
      • layers.py
      • pre_resnet.py
      • resnext_101_64x4d.py
      • __init__.py
      • se_resnext.py
      • wrn_50_2f.py
      • resnext.py
      • .gitignore
      • inceptionresnetv2.py
      • inceptionv4.py
    • cifar10-super-convergence-final94.ipynb
    • dawn_submission
      • fastai_pytorch.json
      • fastai_pytorch.tsv
      • README.md
    • cifar10-topk.ipynb
    • cifar10-super-convergence-more-aug.ipynb
    • jh-darknet.ipynb
    • __init__.py
    • main.py
    • README.md
    • multiproc.py
    • dawn_mod.py
    • .gitignore
    • cifar10-super-convergence-tuned.ipynb
  • LICENSE
  • fp16
    • measure_fp16.ipynb
    • fastai
    • profile_fp16.py
    • profile_fp16_imagenet.py
    • imagenet-fp16.ipynb
    • README.md
    • cifar10-fp16-speed-test.ipynb
  • imagenet_nv
    • jh_warm.py
    • fastai_imagenet.py
    • fp16util.py
    • distributed.py
    • jh_tmp.py
    • models
      • resnext_101_32x4d.py
      • dpn.py
      • se_resnet.py
      • darknet.py
      • resnet.py
      • nasnet.py
      • resnet3.py
      • wideresnet.py
      • resnext_50_32x4d.py
      • convert_torch.py
      • layers.py
      • pre_resnet.py
      • resnext_101_64x4d.py
      • __init__.py
      • resnet2.py
      • se_resnext.py
      • wrn_50_2f.py
      • resnet.py.bak
      • resnext.py
      • .gitignore
      • inceptionresnetv2.py
      • inceptionv4.py
    • blacklist.txt
    • __init__.py
    • main.py
    • README.md
    • multiproc.py
    • blacklist.sh
    • .gitignore
  • README.md
  • scripts
    • resize.py
  • .gitignore
  • aws
    • download_imagenet.ipynb
    • end2end_create_all2.ipynb
    • end2end_demo.ipynb
    • upload_scripts
      • imagenet_formatting.sh
      • train.sh
      • train_nv.sh
      • test.sh
      • download_kaggle_imagenet.sh
      • train_imagenet.sh
      • resize_images.ipynb
      • train2.sh
      • resize_images.py
      • mount_ebs.sh
      • train_cifar10.sh
    • aws_setup.py
    • do.sh
    • run_script.py
    • README.md
    • end2end_create_all.ipynb

End to end ImageNet training

1. Spin up an AWS instance
2. Train ImageNet
3. Save weights and loss

AWS

Check README in AWS directory for instructions on...

1. Creating Spot instances
2. Downloading and formatting ImageNet
3. Running CIFAR10 on an instance

FP16

1. For testing fast.ai model compatability with single precision floating point

ImageNet

Stay tuned!

cifar10

• CIFAR logic
• Dawn submission
• Dawn competition