Skip to content

Latest commit

 

History

History

ptq

README.md

Step-by-Step

This document describes the step-by-step instructions for reproducing PyTorch ResNet50/ResNet18/ResNet101 tuning results with Intel® Low Precision Optimization Tool(LPOT).

Note

  • PyTorch quantization implementation in imperative path has limitation on automatically execution. It requires to manually add QuantStub and DequantStub for quantizable ops, it also requires to manually do fusion operation.
  • LPOT supposes user have done these two steps before invoking LPOT interface. For details, please refer to https://pytorch.org/docs/stable/quantization.html

Prerequisite

1. Installation

pip install -r requirements.txt

2. Prepare Dataset

Download ImageNet Raw image to dir: /path/to/imagenet. The dir include below folder:

ls /path/to/imagenet
train  val

Run

1. ResNet50

cd examples/pytorch/image_recognition/imagenet/cpu/ptq
python main.py -t -a resnet50 --pretrained /path/to/imagenet

2. ResNet18

cd examples/pytorch/image_recognition/imagenet/cpu/ptq
python main.py -t -a resnet18 --pretrained /path/to/imagenet

3. ResNext101_32x8d

cd examples/pytorch/image_recognition/imagenet/cpu/ptq
python main.py -t -a resnext101_32x8d --pretrained /path/to/imagenet

4. InceptionV3

cd examples/pytorch/image_recognition/imagenet/cpu/ptq
python main.py -t -a inception_v3 --pretrained /path/to/imagenet

5. Mobilenet_v2

cd examples/pytorch/image_recognition/imagenet/cpu/ptq
python main.py -t -a mobilenet_v2 --pretrained /path/to/imagenet

6. ResNet50 dump tensors for debug

  cd examples/pytorch/image_recognition/imagenet/cpu/ptq
  python main_dump_tensors.py -t -a resnet50 --pretrained /path/to/imagenet

7. ResNet50 With Intel PyTorch Extension

  cd examples/pytorch/image_recognition/imagenet/cpu/PTQ
  python main.py -t -a resnet50 -j 0 --pretrained --ipex /path/to/imagenet

Saving and loading model:

  • Saving model: After tuning with LPOT, we can get LPOT.model:
from lpot.experimental import Quantization, common
quantizer = Quantization("./conf.yaml")
quantizer.model = common.Model(model)
lpot_model = quantizer()

Here, lpot_model is LPOT model class, so it has "save" API:

lpot_model.save("Path_to_save_configure_file")
  • loading model:
# Without IPEX
model                 # fp32 model
from lpot.utils.pytorch import load
quantized_model = load(
    os.path.join(Path, 'best_configure.yaml'),
    os.path.join(Path, 'best_model_weights.pt'), model)

# With IPEX
import intel_pytorch_extension as ipex 
model                 # fp32 model
model.to(ipex.DEVICE)
try:
    new_model = torch.jit.script(model)
except:
    new_model = torch.jit.trace(model, torch.randn(1, 3, 224, 224).to(ipex.DEVICE))
ipex_config_path = os.path.join(os.path.expanduser(args.tuned_checkpoint),
                                "best_configure.json")
conf = ipex.AmpConf(torch.int8, configure_file=ipex_config_path)
with torch.no_grad():
    for i, (input, target) in enumerate(val_loader):
        with ipex.AutoMixPrecision(conf, running_mode='inference'):
            output = new_model(input.to(ipex.DEVICE))

Please refer to Sample code.

Examples of enabling LPOT auto tuning on PyTorch ResNet

This is a tutorial of how to enable a PyTorch classification model with LPOT.

User Code Analysis

LPOT supports three usages:

  1. User only provide fp32 "model", and configure calibration dataset, evaluation dataset and metric in model-specific yaml config file.
  2. User provide fp32 "model", calibration dataset "q_dataloader" and evaluation dataset "eval_dataloader", and configure metric in tuning.metric field of model-specific yaml config file.
  3. User specifies fp32 "model", calibration dataset "q_dataloader" and a custom "eval_func" which encapsulates the evaluation dataset and metric by itself.

As ResNet18/50/101 series are typical classification models, use Top-K as metric which is built-in supported by LPOT. So here we integrate PyTorch ResNet with LPOT by the first use case for simplicity.

Write Yaml Config File

In examples directory, there is a template.yaml. We could remove most of items and only keep mandotory item for tuning.

model:
  name: imagenet_ptq
  framework: pytorch

quantization:
  calibration:
    sampling_size: 300
    dataloader:
      dataset:
        ImageFolder:
          root: /path/to/calibration/dataset
      transform:
        RandomResizedCrop:
          size: 224
        RandomHorizontalFlip:
        ToTensor:
        Normalize:
          mean: [0.485, 0.456, 0.406]
          std: [0.229, 0.224, 0.225]

evaluation:
  accuracy:
    metric:
      topk: 1
    dataloader:
      batch_size: 30
      dataset:
        ImageFolder:
          root: /path/to/evaluation/dataset
      transform:
        Resize:
          size: 256
        CenterCrop:
          size: 224
        ToTensor:
        Normalize:
          mean: [0.485, 0.456, 0.406]
          std: [0.229, 0.224, 0.225]
  performance:
    configs:
      cores_per_instance: 4
      num_of_instance: 7
    dataloader:
      batch_size: 1
      dataset:
        ImageFolder:
          root: /path/to/evaluation/dataset
      transform:
        Resize:
          size: 256
        CenterCrop:
          size: 224
        ToTensor:
        Normalize:
          mean: [0.485, 0.456, 0.406]
          std: [0.229, 0.224, 0.225]

tuning:
  accuracy_criterion:
    relative:  0.01
  exit_policy:
    timeout: 0
  random_seed: 9527

Here we choose topk built-in metric and set accuracy target as tolerating 0.01 relative accuracy loss of baseline. The default tuning strategy is basic strategy. The timeout 0 means unlimited time for a tuning config meet accuracy target.

Prepare

PyTorch quantization requires two manual steps:

  1. Add QuantStub and DeQuantStub for all quantizable ops.
  2. Fuse possible patterns, such as Conv + Relu and Conv + BN + Relu.

Torchvision provide quantized_model, so we didn't do these steps above for all torchvision models. Please refer torchvision

The related code please refer to examples/pytorch/image_recognition/imagenet/cpu/ptq/main.py.

Code Update

After prepare step is done, we just need update main.py like below.

model.eval()
model.module.fuse_model()
from lpot.experimental import Quantization, common
quantizer = Quantization("./conf.yaml")
quantizer.model = common.Model(model)
q_model = quantizer()

The quantizer() function will return a best quantized model during timeout constrain.

Dump tensors for debug

LPOT can dump every layer output tensor which you specify in evaluation. You just need to add some setting to yaml configure file as below:

tensorboard: true

The default value of "tensorboard" is "off".

For example:

sh run_tuning_dump_tensor.sh --topology=resnet18 --dataset_location=<Dataset>

A "./runs" folder will be generated, for example

ls runs/eval/
tune_0_acc0.73  tune_1_acc0.71 tune_2_acc0.72

"tune_0_acc0.73" means FP32 baseline is accuracy 0.73, and the best tune result is tune_2 with accuracy 0.72. You may want to compare them in tensorboard. It will demonstrate the output tensor and weight of each op in "Histogram", you can also find the tune config of each tuning run in "Text":

tensorboard --bind_all --logdir_spec baseline:./runs/eval/tune_0_acc0.73,tune_2:././runs/eval/tune_2_acc0.72

Tuning With Intel PyTorch Extension

  1. Write Yaml Config File

Add 'backend' field to Yaml Configure and the same for other fields.

  model:
  name: imagenet
  framework: pytorch_ipex
  1. Tuning With LPOT
  from lpot.experimental import Quantization, common
  quantizer = Quantization("./conf_ipex.yaml")
  quantizer.model = common.Model(model)
  lpot_model = quantizer()
  lpot_model.save("Path_to_save_configure_file")
  1. Saving and Run ipex model
  • Saving model
  lpot_model.save("Path_to_save_configure_file")

Here, lpot_model is the result of LPOT tuning. It is LPOT.model class, so it has "save" API.

  • Run ipex model:
import intel_pytorch_extension as ipex 
model                 # fp32 model
model.to(ipex.DEVICE)
try:
    new_model = torch.jit.script(model)
except:
    new_model = torch.jit.trace(model, torch.randn(1, 3, 224, 224).to(ipex.DEVICE))
ipex_config_path = os.path.join(os.path.expanduser(args.tuned_checkpoint),
                                "best_configure.json")
conf = ipex.AmpConf(torch.int8, configure_file=ipex_config_path)
with torch.no_grad():
    for i, (input, target) in enumerate(val_loader):
        with ipex.AutoMixPrecision(conf, running_mode='inference'):
            output = new_model(input.to(ipex.DEVICE))