Examples

Training

Single-instance Training

Code Changes Highlight

There are only a few lines of code change required to use Intel® Extension for PyTorch* on training, as shown:

torch.channels_last should be applied to both of the model object and data to raise CPU resource usage efficiency.
ipex.optimize function applies optimizations against the model object, as well as an optimizer object.

...
import torch
import intel_extension_for_pytorch as ipex
...
model = Model()
model = model.to(memory_format=torch.channels_last)
criterion = ...
optimizer = ...
model.train()
# For Float32
model, optimizer = ipex.optimize(model, optimizer=optimizer)
# For BFloat16
model, optimizer = ipex.optimize(model, optimizer=optimizer, dtype=torch.bfloat16)
...
# Setting memory_format to torch.channels_last could improve performance with 4D input data. This is optional.
data = data.to(memory_format=torch.channels_last)
optimizer.zero_grad()
output = model(data)
...

Complete - Float32

import torch
import torchvision
import intel_extension_for_pytorch as ipex

LR = 0.001
DOWNLOAD = True
DATA = 'datasets/cifar10/'

transform = torchvision.transforms.Compose([
    torchvision.transforms.Resize((224, 224)),
    torchvision.transforms.ToTensor(),
    torchvision.transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))
])
train_dataset = torchvision.datasets.CIFAR10(
        root=DATA,
        train=True,
        transform=transform,
        download=DOWNLOAD,
)
train_loader = torch.utils.data.DataLoader(
        dataset=train_dataset,
        batch_size=128
)

model = torchvision.models.resnet50()
model = model.to(memory_format=torch.channels_last)
criterion = torch.nn.CrossEntropyLoss()
optimizer = torch.optim.SGD(model.parameters(), lr = LR, momentum=0.9)
model.train()
model, optimizer = ipex.optimize(model, optimizer=optimizer)

for batch_idx, (data, target) in enumerate(train_loader):
    # Setting memory_format to torch.channels_last could improve performance with 4D input data. This is optional.
    data = data.to(memory_format=torch.channels_last)
    optimizer.zero_grad()
    output = model(data)
    loss = criterion(output, target)
    loss.backward()
    optimizer.step()
    print(batch_idx)
torch.save({
     'model_state_dict': model.state_dict(),
     'optimizer_state_dict': optimizer.state_dict(),
     }, 'checkpoint.pth')

Complete - BFloat16

import torch
import torchvision
import intel_extension_for_pytorch as ipex

LR = 0.001
DOWNLOAD = True
DATA = 'datasets/cifar10/'

transform = torchvision.transforms.Compose([
    torchvision.transforms.Resize((224, 224)),
    torchvision.transforms.ToTensor(),
    torchvision.transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))
])
train_dataset = torchvision.datasets.CIFAR10(
        root=DATA,
        train=True,
        transform=transform,
        download=DOWNLOAD,
)
train_loader = torch.utils.data.DataLoader(
        dataset=train_dataset,
        batch_size=128
)

model = torchvision.models.resnet50()
model = model.to(memory_format=torch.channels_last)
criterion = torch.nn.CrossEntropyLoss()
optimizer = torch.optim.SGD(model.parameters(), lr = LR, momentum=0.9)
model.train()
model, optimizer = ipex.optimize(model, optimizer=optimizer, dtype=torch.bfloat16)

for batch_idx, (data, target) in enumerate(train_loader):
    optimizer.zero_grad()
    with torch.cpu.amp.autocast():
        # Setting memory_format to torch.channels_last could improve performance with 4D input data. This is optional.
        data = data.to(memory_format=torch.channels_last)
        output = model(data)
        loss = criterion(output, target)
        loss.backward()
    optimizer.step()
    print(batch_idx)
torch.save({
     'model_state_dict': model.state_dict(),
     'optimizer_state_dict': optimizer.state_dict(),
     }, 'checkpoint.pth')

Distributed Training

Distributed training with PyTorch DDP is accelerated by oneAPI Collective Communications Library Bindings for Pytorch* (oneCCL Bindings for Pytorch*). The extension supports FP32 and BF16 data types. More detailed information and examples are available at its Github repo.

Note: When performing distributed training with BF16 data type, use oneCCL Bindings for Pytorch*. Due to a PyTorch limitation, distributed training with BF16 data type with Intel® Extension for PyTorch* is not supported.

import os
import torch
import torch.distributed as dist
import torchvision
import oneccl_bindings_for_pytorch as torch_ccl
import intel_extension_for_pytorch as ipex

LR = 0.001
DOWNLOAD = True
DATA = 'datasets/cifar10/'

transform = torchvision.transforms.Compose([
    torchvision.transforms.Resize((224, 224)),
    torchvision.transforms.ToTensor(),
    torchvision.transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))
])
train_dataset = torchvision.datasets.CIFAR10(
        root=DATA,
        train=True,
        transform=transform,
        download=DOWNLOAD,
)
train_loader = torch.utils.data.DataLoader(
        dataset=train_dataset,
        batch_size=128
)

os.environ['MASTER_ADDR'] = '127.0.0.1'
os.environ['MASTER_PORT'] = '29500'
os.environ['RANK'] = os.environ.get('PMI_RANK', 0)
os.environ['WORLD_SIZE'] = os.environ.get('PMI_SIZE', 1)
dist.init_process_group(
backend='ccl',
init_method='env://'
)

model = torchvision.models.resnet50()
criterion = torch.nn.CrossEntropyLoss()
optimizer = torch.optim.SGD(model.parameters(), lr = LR, momentum=0.9)
model.train()
model, optimizer = ipex.optimize(model, optimizer=optimizer)

model = torch.nn.parallel.DistributedDataParallel(model)

for batch_idx, (data, target) in enumerate(train_loader):
    optimizer.zero_grad()
    # Setting memory_format to torch.channels_last could improve performance with 4D input data. This is optional.
    data = data.to(memory_format=torch.channels_last)
    output = model(data)
    loss = criterion(output, target)
    loss.backward()
    optimizer.step()
    print('batch_id: {}'.format(batch_idx))
torch.save({
     'model_state_dict': model.state_dict(),
     'optimizer_state_dict': optimizer.state_dict(),
     }, 'checkpoint.pth')

Inference

Channels last is a memory layout format that is more friendly to Intel Architecture. We recommend using this memory layout format for computer vision workloads by invoking to(memory_format=torch.channels_last) function against the model object and input data.

The optimize function of Intel® Extension for PyTorch* applies optimizations to the model, bringing additional performance boosts. For both computer vision workloads and NLP workloads, we recommend applying the optimize function against the model object.

Float32

Imperative Mode

Resnet50

import torch
import torchvision.models as models

model = models.resnet50(pretrained=True)
model.eval()
data = torch.rand(1, 3, 224, 224)

model = model.to(memory_format=torch.channels_last)
data = data.to(memory_format=torch.channels_last)

#################### code changes ####################
import intel_extension_for_pytorch as ipex
model = ipex.optimize(model)
######################################################

with torch.no_grad():
  model(data)

BERT

import torch
from transformers import BertModel

model = BertModel.from_pretrained(args.model_name)
model.eval()

vocab_size = model.config.vocab_size
batch_size = 1
seq_length = 512
data = torch.randint(vocab_size, size=[batch_size, seq_length])

#################### code changes ####################
import intel_extension_for_pytorch as ipex
model = ipex.optimize(model)
######################################################

with torch.no_grad():
  model(data)

TorchScript Mode

We recommend you take advantage of Intel® Extension for PyTorch* with TorchScript for further optimizations.

Resnet50

import torch
import torchvision.models as models

model = models.resnet50(pretrained=True)
model.eval()
data = torch.rand(1, 3, 224, 224)

model = model.to(memory_format=torch.channels_last)
data = data.to(memory_format=torch.channels_last)

#################### code changes ####################
import intel_extension_for_pytorch as ipex
model = ipex.optimize(model)
######################################################

with torch.no_grad():
  d = torch.rand(1, 3, 224, 224)
  model = torch.jit.trace(model, d)
  model = torch.jit.freeze(model)

  model(data)

BERT

import torch
from transformers import BertModel

model = BertModel.from_pretrained(args.model_name)
model.eval()

vocab_size = model.config.vocab_size
batch_size = 1
seq_length = 512
data = torch.randint(vocab_size, size=[batch_size, seq_length])

#################### code changes ####################
import intel_extension_for_pytorch as ipex
model = ipex.optimize(model)
######################################################

with torch.no_grad():
  d = torch.randint(vocab_size, size=[batch_size, seq_length])
  model = torch.jit.trace(model, (d,), check_trace=False, strict=False)
  model = torch.jit.freeze(model)

  model(data)

BFloat16

Similar to running with FP32, the optimize function also works for BFloat16 data type. The only difference is setting dtype parameter to torch.bfloat16. We recommend using Auto Mixed Precision (AMP) with BFloat16 data type.

Imperative Mode

Resnet50

import torch
import torchvision.models as models

model = models.resnet50(pretrained=True)
model.eval()
data = torch.rand(1, 3, 224, 224)

model = model.to(memory_format=torch.channels_last)
data = data.to(memory_format=torch.channels_last)

#################### code changes ####################
import intel_extension_for_pytorch as ipex
model = ipex.optimize(model, dtype=torch.bfloat16)
######################################################

with torch.no_grad():
  with torch.cpu.amp.autocast():
    model(data)

BERT

import torch
from transformers import BertModel

model = BertModel.from_pretrained(args.model_name)
model.eval()

vocab_size = model.config.vocab_size
batch_size = 1
seq_length = 512
data = torch.randint(vocab_size, size=[batch_size, seq_length])

#################### code changes ####################
import intel_extension_for_pytorch as ipex
model = ipex.optimize(model, dtype=torch.bfloat16)
######################################################

with torch.no_grad():
  with torch.cpu.amp.autocast():
    model(data)

TorchScript Mode

We recommend you take advantage of Intel® Extension for PyTorch* with TorchScript for further optimizations.

Resnet50

import torch
import torchvision.models as models

model = models.resnet50(pretrained=True)
model.eval()
data = torch.rand(1, 3, 224, 224)

model = model.to(memory_format=torch.channels_last)
data = data.to(memory_format=torch.channels_last)

#################### code changes ####################
import intel_extension_for_pytorch as ipex
model = ipex.optimize(model, dtype=torch.bfloat16)
######################################################

with torch.no_grad():
  with torch.cpu.amp.autocast():
    model = torch.jit.trace(model, torch.rand(1, 3, 224, 224))
    model = torch.jit.freeze(model)

    model(data)

BERT

import torch
from transformers import BertModel

model = BertModel.from_pretrained(args.model_name)
model.eval()

vocab_size = model.config.vocab_size
batch_size = 1
seq_length = 512
data = torch.randint(vocab_size, size=[batch_size, seq_length])

#################### code changes ####################
import intel_extension_for_pytorch as ipex
model = ipex.optimize(model, dtype=torch.bfloat16)
######################################################

with torch.no_grad():
  with torch.cpu.amp.autocast():
    d = torch.randint(vocab_size, size=[batch_size, seq_length])
    model = torch.jit.trace(model, (d,), check_trace=False, strict=False)
    model = torch.jit.freeze(model)

    model(data)

INT8

Starting from Intel® Extension for PyTorch* 1.12.0, quantization feature supports both static and dynamic modes.

Calibration

Static Quantization

Please follow the steps below to perform static calibration:

Import intel_extension_for_pytorch as ipex.
Import prepare and convert from intel_extension_for_pytorch.quantization.
Instantiate a config object from torch.ao.quantization.QConfig to save configuration data during calibration.
Prepare model for calibration.
Perform calibration against dataset.
Invoke ipex.quantization.convert function to apply the calibration configure object to the fp32 model object to get an INT8 model.
Save the INT8 model into a pt file.

import os
import torch
#################### code changes ####################
import intel_extension_for_pytorch as ipex
from intel_extension_for_pytorch.quantization import prepare, convert
######################################################

model = Model()
model.eval()
data = torch.rand(<shape>)

qconfig = ipex.quantization.default_static_qconfig
# Alternatively, define your own qconfig:
#from torch.ao.quantization import MinMaxObserver, PerChannelMinMaxObserver, QConfig
#qconfig = QConfig(activation=MinMaxObserver.with_args(qscheme=torch.per_tensor_affine, dtype=torch.quint8),
#        weight=PerChannelMinMaxObserver.with_args(dtype=torch.qint8, qscheme=torch.per_channel_symmetric))
prepared_model = prepare(model, qconfig, example_inputs=data, inplace=False)

for d in calibration_data_loader():
  prepared_model(d)

converted_model = convert(prepared_model)
with torch.no_grad():
  traced_model = torch.jit.trace(converted_model, data)
  traced_model = torch.jit.freeze(traced_model)

traced_model.save("quantized_model.pt")

Dynamic Quantization

Please follow the steps below to perform static calibration:

Import intel_extension_for_pytorch as ipex.
Import prepare and convert from intel_extension_for_pytorch.quantization.
Instantiate a config object from torch.ao.quantization.QConfig to save configuration data during calibration.
Prepare model for quantization.
Convert the model.
Run inference to perform dynamic quantization.
Save the INT8 model into a pt file.

import os
import torch
#################### code changes ####################
import intel_extension_for_pytorch as ipex
from intel_extension_for_pytorch.quantization import prepare, convert
######################################################

model = Model()
model.eval()
data = torch.rand(<shape>)

dynamic_qconfig = ipex.quantization.default_dynamic_qconfig
# Alternatively, define your own qconfig:
#from torch.ao.quantization import MinMaxObserver, PlaceholderObserver, QConfig
#qconfig = QConfig(
#        activation = PlaceholderObserver.with_args(dtype=torch.float, compute_dtype=torch.quint8),
#        weight = PerChannelMinMaxObserver.with_args(dtype=torch.qint8, qscheme=torch.per_channel_symmetric))
prepared_model = prepare(model, qconfig, example_inputs=data)

converted_model = convert(prepared_model)
with torch.no_grad():
  traced_model = torch.jit.trace(converted_model, data)
  traced_model = torch.jit.freeze(traced_model)

traced_model.save("quantized_model.pt")

Deployment

For deployment, the INT8 model is loaded from the local file and can be used directly on the inference.

Follow the steps below:

Import intel_extension_for_pytorch as ipex.
Load the INT8 model from the saved file.
Run inference.

import torch
#################### code changes ####################
import intel_extension_for_pytorch as ipex
######################################################

model = torch.jit.load('quantization_model.pt')
model.eval()
model = torch.jit.freeze(model)
data = torch.rand(<shape>)

with torch.no_grad():
  model(data)

oneDNN provides oneDNN Graph Compiler as a prototype feature that could boost performance for selective topologies. No code change is required. Install a binary with this feature enabled. We verified this feature with Bert-large, bert-base-cased, roberta-base, xlm-roberta-base, google-electra-base-generator and google-electra-base-discriminator.

C++

To work with libtorch, C++ library of PyTorch, Intel® Extension for PyTorch* provides its C++ dynamic library as well. The C++ library is supposed to handle inference workload only, such as service deployment. For regular development, use the Python interface. Unlike using libtorch, no specific code changes are required, except for converting input data into channels last data format. Compilation follows the recommended methodology with CMake. Detailed instructions can be found in PyTorch tutorial.

During compilation, Intel optimizations will be activated automatically once C++ dynamic library of Intel® Extension for PyTorch* is linked.

The example code below works for all data types.

example-app.cpp

#include <torch/script.h>
#include <iostream>
#include <memory>

int main(int argc, const char* argv[]) {
    torch::jit::script::Module module;
    try {
        module = torch::jit::load(argv[1]);
    }
    catch (const c10::Error& e) {
        std::cerr << "error loading the model\n";
        return -1;
    }
    std::vector<torch::jit::IValue> inputs;
    // make sure input data are converted to channels last format
    inputs.push_back(torch::ones({1, 3, 224, 224}).to(c10::MemoryFormat::ChannelsLast));

    at::Tensor output = module.forward(inputs).toTensor();

    return 0;
}

CMakeLists.txt

cmake_minimum_required(VERSION 3.0 FATAL_ERROR)
project(example-app)

find_package(intel_ext_pt_cpu REQUIRED)

add_executable(example-app example-app.cpp)
target_link_libraries(example-app "${TORCH_LIBRARIES}")

set_property(TARGET example-app PROPERTY CXX_STANDARD 14)

Command for compilation

$ cmake -DCMAKE_PREFIX_PATH=<LIBPYTORCH_PATH> ..
$ make

If Found INTEL_EXT_PT_CPU is shown as TRUE, the extension had been linked into the binary. This can be verified with Linux command ldd.

$ cmake -DCMAKE_PREFIX_PATH=/workspace/libtorch ..
-- The C compiler identification is GNU 9.3.0
-- The CXX compiler identification is GNU 9.3.0
-- Check for working C compiler: /usr/bin/cc
-- Check for working C compiler: /usr/bin/cc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Check for working CXX compiler: /usr/bin/c++
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Looking for pthread.h
-- Looking for pthread.h - found
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD
-- Performing Test CMAKE_HAVE_LIBC_PTHREAD - Failed
-- Looking for pthread_create in pthreads
-- Looking for pthread_create in pthreads - not found
-- Looking for pthread_create in pthread
-- Looking for pthread_create in pthread - found
-- Found Threads: TRUE
-- Found Torch: /workspace/libtorch/lib/libtorch.so
-- Found INTEL_EXT_PT_CPU: TRUE
-- Configuring done
-- Generating done
-- Build files have been written to: /workspace/build

$ ldd example-app
        ...
        libtorch.so => /workspace/libtorch/lib/libtorch.so (0x00007f3cf98e0000)
        libc10.so => /workspace/libtorch/lib/libc10.so (0x00007f3cf985a000)
        libintel-ext-pt-cpu.so => /workspace/libtorch/lib/libintel-ext-pt-cpu.so (0x00007f3cf70fc000)
        libtorch_cpu.so => /workspace/libtorch/lib/libtorch_cpu.so (0x00007f3ce16ac000)
        ...
        libdnnl_graph.so.0 => /workspace/libtorch/lib/libdnnl_graph.so.0 (0x00007f3cde954000)
        ...

Model Zoo

Use cases that had already been optimized by Intel engineers are available at Model Zoo for Intel® Architecture. A bunch of PyTorch use cases for benchmarking are also available on the GitHub page. You can get performance benefits out-of-box by simply running scipts in the Model Zoo.