Home

Awesome

Introduction

This repository holds NVIDIA-maintained utilities to streamline mixed precision and distributed training in Pytorch. Some of the code here will be included in upstream Pytorch eventually. The intent of Apex is to make up-to-date utilities available to users as quickly as possible.

Full API Documentation: https://nvidia.github.io/apex

GTC 2019 and Pytorch DevCon 2019 Slides

Contents

1. Amp: Automatic Mixed Precision

Deprecated. Use PyTorch AMP

apex.amp is a tool to enable mixed precision training by changing only 3 lines of your script. Users can easily experiment with different pure and mixed precision training modes by supplying different flags to amp.initialize.

Webinar introducing Amp (The flag cast_batchnorm has been renamed to keep_batchnorm_fp32).

API Documentation

Comprehensive Imagenet example

DCGAN example coming soon...

Moving to the new Amp API (for users of the deprecated "Amp" and "FP16_Optimizer" APIs)

2. Distributed Training

apex.parallel.DistributedDataParallel is deprecated. Use torch.nn.parallel.DistributedDataParallel

apex.parallel.DistributedDataParallel is a module wrapper, similar to torch.nn.parallel.DistributedDataParallel. It enables convenient multiprocess distributed training, optimized for NVIDIA's NCCL communication library.

API Documentation

Python Source

Example/Walkthrough

The Imagenet example shows use of apex.parallel.DistributedDataParallel along with apex.amp.

Synchronized Batch Normalization

Deprecated. Use torch.nn.SyncBatchNorm

apex.parallel.SyncBatchNorm extends torch.nn.modules.batchnorm._BatchNorm to support synchronized BN. It allreduces stats across processes during multiprocess (DistributedDataParallel) training. Synchronous BN has been used in cases where only a small local minibatch can fit on each GPU. Allreduced stats increase the effective batch size for the BN layer to the global batch size across all processes (which, technically, is the correct formulation). Synchronous BN has been observed to improve converged accuracy in some of our research models.

Checkpointing

To properly save and load your amp training, we introduce the amp.state_dict(), which contains all loss_scalers and their corresponding unskipped steps, as well as amp.load_state_dict() to restore these attributes.

In order to get bitwise accuracy, we recommend the following workflow:

# Initialization
opt_level = 'O1'
model, optimizer = amp.initialize(model, optimizer, opt_level=opt_level)

# Train your model
...
with amp.scale_loss(loss, optimizer) as scaled_loss:
    scaled_loss.backward()
...

# Save checkpoint
checkpoint = {
    'model': model.state_dict(),
    'optimizer': optimizer.state_dict(),
    'amp': amp.state_dict()
}
torch.save(checkpoint, 'amp_checkpoint.pt')
...

# Restore
model = ...
optimizer = ...
checkpoint = torch.load('amp_checkpoint.pt')

model, optimizer = amp.initialize(model, optimizer, opt_level=opt_level)
model.load_state_dict(checkpoint['model'])
optimizer.load_state_dict(checkpoint['optimizer'])
amp.load_state_dict(checkpoint['amp'])

# Continue training
...

Note that we recommend restoring the model using the same opt_level. Also note that we recommend calling the load_state_dict methods after amp.initialize.

Installation

Each apex.contrib module requires one or more install options other than --cpp_ext and --cuda_ext. Note that contrib modules do not necessarily support stable PyTorch releases.

Containers

NVIDIA PyTorch Containers are available on NGC: https://catalog.ngc.nvidia.com/orgs/nvidia/containers/pytorch. The containers come with all the custom extensions available at the moment.

See the NGC documentation for details such as:

From Source

To install Apex from source, we recommend using the nightly Pytorch obtainable from https://github.com/pytorch/pytorch.

The latest stable release obtainable from https://pytorch.org should also work.

We recommend installing Ninja to make compilation faster.

Linux

For performance and full functionality, we recommend installing Apex with CUDA and C++ extensions via

git clone https://github.com/NVIDIA/apex
cd apex
# if pip >= 23.1 (ref: https://pip.pypa.io/en/stable/news/#v23-1) which supports multiple `--config-settings` with the same key... 
pip install -v --disable-pip-version-check --no-cache-dir --no-build-isolation --config-settings "--build-option=--cpp_ext" --config-settings "--build-option=--cuda_ext" ./
# otherwise
pip install -v --disable-pip-version-check --no-cache-dir --no-build-isolation --global-option="--cpp_ext" --global-option="--cuda_ext" ./

APEX also supports a Python-only build via

pip install -v --disable-pip-version-check --no-build-isolation --no-cache-dir ./

A Python-only build omits:

[Experimental] Windows

pip install -v --disable-pip-version-check --no-cache-dir --no-build-isolation --config-settings "--build-option=--cpp_ext" --config-settings "--build-option=--cuda_ext" . may work if you were able to build Pytorch from source on your system. A Python-only build via pip install -v --no-cache-dir . is more likely to work.
If you installed Pytorch in a Conda environment, make sure to install Apex in that same environment.

Custom C++/CUDA Extensions and Install Options

If a requirement of a module is not met, then it will not be built.

Module NameInstall OptionMisc
apex_C--cpp_ext
amp_C--cuda_ext
syncbn--cuda_ext
fused_layer_norm_cuda--cuda_extapex.normalization
mlp_cuda--cuda_ext
scaled_upper_triang_masked_softmax_cuda--cuda_ext
generic_scaled_masked_softmax_cuda--cuda_ext
scaled_masked_softmax_cuda--cuda_ext
fused_weight_gradient_mlp_cuda--cuda_extRequires CUDA>=11
permutation_search_cuda--permutation_searchapex.contrib.sparsity
bnp--bnpapex.contrib.groupbn
xentropy--xentropyapex.contrib.xentropy
focal_loss_cuda--focal_lossapex.contrib.focal_loss
fused_index_mul_2d--index_mul_2dapex.contrib.index_mul_2d
fused_adam_cuda--deprecated_fused_adamapex.contrib.optimizers
fused_lamb_cuda--deprecated_fused_lambapex.contrib.optimizers
fast_layer_norm--fast_layer_normapex.contrib.layer_norm. different from fused_layer_norm
fmhalib--fmhaapex.contrib.fmha
fast_multihead_attn--fast_multihead_attnapex.contrib.multihead_attn
transducer_joint_cuda--transducerapex.contrib.transducer
transducer_loss_cuda--transducerapex.contrib.transducer
cudnn_gbn_lib--cudnn_gbnRequires cuDNN>=8.5, apex.contrib.cudnn_gbn
peer_memory_cuda--peer_memoryapex.contrib.peer_memory
nccl_p2p_cuda--nccl_p2pRequires NCCL >= 2.10, apex.contrib.nccl_p2p
fast_bottleneck--fast_bottleneckRequires peer_memory_cuda and nccl_p2p_cuda, apex.contrib.bottleneck
fused_conv_bias_relu--fused_conv_bias_reluRequires cuDNN>=8.4, apex.contrib.conv_bias_relu