Convert and Optimize YOLOv8 keypoint detection model with OpenVINO™

This Jupyter notebook can be launched on-line, opening an interactive environment in a browser window. You can also make a local installation. Choose one of the following options:

Keypoint detection/Pose is a task that involves detecting specific points in an image or video frame. These points are referred to as keypoints and are used to track movement or pose estimation. YOLOv8 can detect keypoints in an image or video frame with high accuracy and speed.

This tutorial demonstrates step-by-step instructions on how to run and optimize PyTorch YOLOv8 Pose model with OpenVINO. We consider the steps required for keypoint detection scenario.

The tutorial consists of the following steps: - Prepare the PyTorch model. - Download and prepare a dataset. - Validate the original model. - Convert the PyTorch model to OpenVINO IR. - Validate the converted model. - Prepare and run optimization pipeline. - Compare performance of the FP32 and quantized models. - Compare accuracy of the FP32 and quantized models. - Live demo

Table of contents:

• Get PyTorch model

  • Prerequisites

• Instantiate model

  • Convert model to OpenVINO IR

  • Verify model inference

  • Select inference device

  • Test on single image

• Check model accuracy on the dataset

  • Download the validation dataset

  • Define validation function

  • Configure Validator helper and create DataLoader

• Optimize model using NNCF Post-training Quantization API

  • Validate Quantized model inference

• Compare the Original and Quantized Models

  • Compare performance of the Original and Quantized Models

  • Compare accuracy of the Original and Quantized Models

• Other ways to optimize model

• Live demo

  • Run Keypoint Detection on video

Installation Instructions

This is a self-contained example that relies solely on its own code.

We recommend running the notebook in a virtual environment. You only need a Jupyter server to start. For details, please refer to Installation Guide.

Get PyTorch model

Generally, PyTorch models represent an instance of the torch.nn.Module class, initialized by a state dictionary with model weights. We will use the YOLOv8 nano model (also known as yolov8n) pre-trained on a COCO dataset, which is available in this repo. Similar steps are also applicable to other YOLOv8 models. Typical steps to obtain a pre-trained model: 1. Create an instance of a model class. 2. Load a checkpoint state dict, which contains the pre-trained model weights. 3. Turn the model to evaluation for switching some operations to inference mode.

In this case, the creators of the model provide an API that enables converting the YOLOv8 model to OpenVINO IR. Therefore, we do not need to do these steps manually.

Prerequisites

Install necessary packages.

%pip install -q "openvino>=2024.0.0" "nncf>=2.9.0"
%pip install -q "protobuf==3.20.*" "torch>=2.1" "torchvision>=0.16" "ultralytics==8.2.24" "onnx" tqdm opencv-python --extra-index-url https://download.pytorch.org/whl/cpu

Import required utility functions. The lower cell will download the notebook_utils Python module from GitHub.

from pathlib import Path

# Fetch `notebook_utils` module
import requests

r = requests.get(
    url="https://raw.githubusercontent.com/openvinotoolkit/openvino_notebooks/latest/utils/notebook_utils.py",
)

open("notebook_utils.py", "w").write(r.text)
from notebook_utils import download_file, VideoPlayer, device_widget

# Download a test sample
IMAGE_PATH = Path("./data/intel_rnb.jpg")
download_file(
    url="https://storage.openvinotoolkit.org/repositories/openvino_notebooks/data/data/image/intel_rnb.jpg",
    filename=IMAGE_PATH.name,
    directory=IMAGE_PATH.parent,
)

data/intel_rnb.jpg:   0%|          | 0.00/288k [00:00<?, ?B/s]

PosixPath('/home/akash/intel/openvino_notebooks/notebooks/yolov8-optimization/data/intel_rnb.jpg')

Instantiate model

For loading the model, required to specify a path to the model checkpoint. It can be some local path or name available on models hub (in this case model checkpoint will be downloaded automatically).

Making prediction, the model accepts a path to input image and returns list with Results class object. Results contains boxes and key points. Also it contains utilities for processing results, for example, plot() method for drawing.

Let us consider the examples:

models_dir = Path("./models")
models_dir.mkdir(exist_ok=True)

from PIL import Image
from ultralytics import YOLO

POSE_MODEL_NAME = "yolov8n-pose"

pose_model = YOLO(models_dir / f"{POSE_MODEL_NAME}.pt")
label_map = pose_model.model.names

res = pose_model(IMAGE_PATH)
Image.fromarray(res[0].plot()[:, :, ::-1])

Downloading ultralytics/assets to 'models/yolov8n-pose.pt'...

100%|██████████████████████████████████████████████████████████████████████████████████████████████████████| 6.52M/6.52M [00:02<00:00, 3.06MB/s]

image 1/1 /home/akash/intel/openvino_notebooks/notebooks/yolov8-optimization/data/intel_rnb.jpg: 480x640 1 person, 88.5ms
Speed: 2.4ms preprocess, 88.5ms inference, 480.4ms postprocess per image at shape (1, 3, 480, 640)

Convert model to OpenVINO IR

YOLOv8 provides API for convenient model exporting to different formats including OpenVINO IR. model.export is responsible for model conversion. We need to specify the format, and additionally, we can preserve dynamic shapes in the model.

# object detection model
pose_model_path = models_dir / f"{POSE_MODEL_NAME}_openvino_model/{POSE_MODEL_NAME}.xml"
if not pose_model_path.exists():
    pose_model.export(format="openvino", dynamic=True, half=True)

Ultralytics YOLOv8.2.24 🚀 Python-3.8.10 torch-2.1.0+cu121 CPU (Intel Core(TM) i9-10980XE 3.00GHz)

PyTorch: starting from 'models/yolov8n-pose.pt' with input shape (1, 3, 640, 640) BCHW and output shape(s) (1, 56, 8400) (6.5 MB)

OpenVINO: starting export with openvino 2024.3.0-16041-1e3b88e4e3f-releases/2024/3...
OpenVINO: export success ✅ 1.9s, saved as 'models/yolov8n-pose_openvino_model/' (6.7 MB)

Export complete (3.4s)
Results saved to /home/akash/intel/openvino_notebooks/notebooks/yolov8-optimization/models
Predict:         yolo predict task=pose model=models/yolov8n-pose_openvino_model imgsz=640 half
Validate:        yolo val task=pose model=models/yolov8n-pose_openvino_model imgsz=640 data=/usr/src/app/ultralytics/datasets/coco-pose.yaml half
Visualize:       https://netron.app

Verify model inference

We can reuse the base model pipeline for pre- and postprocessing just replacing the inference method where we will use the IR model for inference.

Select inference device

Select device from dropdown list for running inference using OpenVINO

device = device_widget()

device

Dropdown(description='Device:', index=1, options=('CPU', 'AUTO'), value='AUTO')

Test on single image

Now, once we have defined preprocessing and postprocessing steps, we are ready to check model prediction.

import torch
import openvino as ov

core = ov.Core()
pose_ov_model = core.read_model(pose_model_path)

ov_config = {}
if device.value != "CPU":
    pose_ov_model.reshape({0: [1, 3, 640, 640]})
if "GPU" in device.value or ("AUTO" in device.value and "GPU" in core.available_devices):
    ov_config = {"GPU_DISABLE_WINOGRAD_CONVOLUTION": "YES"}
pose_compiled_model = core.compile_model(pose_ov_model, device.value, ov_config)


def infer(*args):
    result = pose_compiled_model(args)
    return torch.from_numpy(result[0])


pose_model.predictor.inference = infer
pose_model.predictor.model.pt = False

res = pose_model(IMAGE_PATH)
Image.fromarray(res[0].plot()[:, :, ::-1])

image 1/1 /home/akash/intel/openvino_notebooks/notebooks/yolov8-optimization/data/intel_rnb.jpg: 640x640 2 persons, 16.7ms
Speed: 4.9ms preprocess, 16.7ms inference, 1.2ms postprocess per image at shape (1, 3, 640, 640)

Great! The result is the same, as produced by original models.

Check model accuracy on the dataset

For comparing the optimized model result with the original, it is good to know some measurable results in terms of model accuracy on the validation dataset.

Download the validation dataset

YOLOv8 is pre-trained on the COCO dataset, so to evaluate the model accuracy we need to download it. According to the instructions provided in the YOLOv8 repo, we also need to download annotations in the format used by the author of the model, for use with the original model evaluation function.

Note: The initial dataset download may take a few minutes to complete. The download speed will vary depending on the quality of your internet connection.

from zipfile import ZipFile

from ultralytics.data.utils import DATASETS_DIR


DATA_URL = "https://ultralytics.com/assets/coco8-pose.zip"
CFG_URL = "https://raw.githubusercontent.com/ultralytics/ultralytics/v8.1.0/ultralytics/cfg/datasets/coco8-pose.yaml"

OUT_DIR = DATASETS_DIR

DATA_PATH = OUT_DIR / "val2017.zip"
CFG_PATH = OUT_DIR / "coco8-pose.yaml"

download_file(DATA_URL, DATA_PATH.name, DATA_PATH.parent)
download_file(CFG_URL, CFG_PATH.name, CFG_PATH.parent)

if not (OUT_DIR / "coco8-pose/labels").exists():
    with ZipFile(DATA_PATH, "r") as zip_ref:
        zip_ref.extractall(OUT_DIR)

/home/akash/intel/NNCF/nncf/examples/post_training_quantization/openvino/yolov8/datasets/val2017.zip:   0%|   …

/home/akash/intel/NNCF/nncf/examples/post_training_quantization/openvino/yolov8/datasets/coco8-pose.yaml:   0%…

Define validation function

import numpy as np
from tqdm.notebook import tqdm
from ultralytics.utils.metrics import ConfusionMatrix


def test(
    model: ov.Model,
    core: ov.Core,
    data_loader: torch.utils.data.DataLoader,
    validator,
    num_samples: int = None,
):
    """
    OpenVINO YOLOv8 model accuracy validation function. Runs model validation on dataset and returns metrics
    Parameters:
        model (Model): OpenVINO model
        data_loader (torch.utils.data.DataLoader): dataset loader
        validator: instance of validator class
        num_samples (int, *optional*, None): validate model only on specified number samples, if provided
    Returns:
        stats: (Dict[str, float]) - dictionary with aggregated accuracy metrics statistics, key is metric name, value is metric value
    """
    validator.seen = 0
    validator.jdict = []
    validator.stats = dict(tp_p=[], tp=[], conf=[], pred_cls=[], target_cls=[])
    validator.batch_i = 1
    validator.confusion_matrix = ConfusionMatrix(nc=validator.nc)
    model.reshape({0: [1, 3, -1, -1]})
    compiled_model = core.compile_model(model)
    for batch_i, batch in enumerate(tqdm(data_loader, total=num_samples)):
        if num_samples is not None and batch_i == num_samples:
            break
        batch = validator.preprocess(batch)
        results = compiled_model(batch["img"])
        preds = torch.from_numpy(results[compiled_model.output(0)])
        preds = validator.postprocess(preds)
        validator.update_metrics(preds, batch)
    stats = validator.get_stats()
    return stats


def print_stats(stats: np.ndarray, total_images: int, total_objects: int):
    """
    Helper function for printing accuracy statistic
    Parameters:
        stats: (Dict[str, float]) - dictionary with aggregated accuracy metrics statistics, key is metric name, value is metric value
        total_images (int) -  number of evaluated images
        total objects (int)
    Returns:
        None
    """
    print("Boxes:")
    mp, mr, map50, mean_ap = (
        stats["metrics/precision(B)"],
        stats["metrics/recall(B)"],
        stats["metrics/mAP50(B)"],
        stats["metrics/mAP50-95(B)"],
    )
    # Print results
    print("    Best mean average:")
    s = ("%20s" + "%12s" * 6) % (
        "Class",
        "Images",
        "Labels",
        "Precision",
        "Recall",
        "mAP@.5",
        "mAP@.5:.95",
    )
    print(s)
    pf = "%20s" + "%12i" * 2 + "%12.3g" * 4  # print format
    print(pf % ("all", total_images, total_objects, mp, mr, map50, mean_ap))
    if "metrics/precision(M)" in stats:
        s_mp, s_mr, s_map50, s_mean_ap = (
            stats["metrics/precision(M)"],
            stats["metrics/recall(M)"],
            stats["metrics/mAP50(M)"],
            stats["metrics/mAP50-95(M)"],
        )
        # Print results
        print("    Macro average mean:")
        s = ("%20s" + "%12s" * 6) % (
            "Class",
            "Images",
            "Labels",
            "Precision",
            "Recall",
            "mAP@.5",
            "mAP@.5:.95",
        )
        print(s)
        pf = "%20s" + "%12i" * 2 + "%12.3g" * 4  # print format
        print(pf % ("all", total_images, total_objects, s_mp, s_mr, s_map50, s_mean_ap))

Configure Validator helper and create DataLoader

The original model repository uses a Validator wrapper, which represents the accuracy validation pipeline. It creates dataloader and evaluation metrics and updates metrics on each data batch produced by the dataloader. Besides that, it is responsible for data preprocessing and results postprocessing. For class initialization, the configuration should be provided. We will use the default setup, but it can be replaced with some parameters overriding to test on custom data. The model has connected the ValidatorClass method, which creates a validator class instance.

from ultralytics.utils import DEFAULT_CFG
from ultralytics.cfg import get_cfg
from ultralytics.data.utils import check_det_dataset

args = get_cfg(cfg=DEFAULT_CFG)
args.data = "coco8-pose.yaml"
args.model = "yolov8n-pose.pt"

from ultralytics.models.yolo.pose import PoseValidator

pose_validator = PoseValidator(args=args)

pose_validator.data = check_det_dataset(args.data)
pose_validator.stride = 32
pose_data_loader = pose_validator.get_dataloader(OUT_DIR / "coco8-pose", 1)

val: Scanning /home/akash/intel/NNCF/nncf/examples/post_training_quantization/openvino/yolov8/datasets/coco8-pose/labels/train.cache... 8 images

from ultralytics.utils.metrics import OKS_SIGMA

pose_validator.is_coco = True
pose_validator.names = pose_model.model.names
pose_validator.metrics.names = pose_validator.names
pose_validator.nc = pose_model.model.model[-1].nc
pose_validator.sigma = OKS_SIGMA

After definition test function and validator creation, we are ready for getting accuracy metrics >Note: Model evaluation is time consuming process and can take several minutes, depending on the hardware. For reducing calculation time, we define num_samples parameter with evaluation subset size, but in this case, accuracy can be noncomparable with originally reported by the authors of the model, due to validation subset difference. To validate the models on the full dataset set ``NUM_TEST_SAMPLES = None``.

NUM_TEST_SAMPLES = 300

fp_pose_stats = test(pose_ov_model, core, pose_data_loader, pose_validator, num_samples=NUM_TEST_SAMPLES)

0%|          | 0/300 [00:00<?, ?it/s]

print_stats(fp_pose_stats, pose_validator.seen, pose_validator.nt_per_class.sum())

Boxes:
    Best mean average:
               Class      Images      Labels   Precision      Recall      mAP@.5  mAP@.5:.95
                 all           8          21           1       0.899       0.955       0.736

print_stats reports the following list of accuracy metrics:

• Precision is the degree of exactness of the model in identifying only relevant objects.

• Recall measures the ability of the model to detect all ground truths objects.

• mAP@t - mean average precision, represented as area under the Precision-Recall curve aggregated over all classes in the dataset, where t is the Intersection Over Union (IOU) threshold, degree of overlapping between ground truth and predicted objects. Therefore, mAP@.5 indicates that mean average precision is calculated at 0.5 IOU threshold, mAP@.5:.95 - is calculated on range IOU thresholds from 0.5 to 0.95 with step 0.05.

Optimize model using NNCF Post-training Quantization API

NNCF provides a suite of advanced algorithms for Neural Networks inference optimization in OpenVINO with minimal accuracy drop. We will use 8-bit quantization in post-training mode (without the fine-tuning pipeline) to optimize YOLOv8.

The optimization process contains the following steps:

1. Create a Dataset for quantization.

2. Run nncf.quantize for getting an optimized model.

3. Serialize OpenVINO IR model, using the openvino.runtime.serialize function.

Please select below whether you would like to run quantization to improve model inference speed.

import ipywidgets as widgets

int8_model_pose_path = models_dir / f"{POSE_MODEL_NAME}_openvino_int8_model/{POSE_MODEL_NAME}.xml"

to_quantize = widgets.Checkbox(
    value=True,
    description="Quantization",
    disabled=False,
)

to_quantize

Checkbox(value=True, description='Quantization')

Let’s load skip magic extension to skip quantization if to_quantize is not selected

# Fetch skip_kernel_extension module
r = requests.get(
    url="https://raw.githubusercontent.com/openvinotoolkit/openvino_notebooks/latest/utils/skip_kernel_extension.py",
)
open("skip_kernel_extension.py", "w").write(r.text)

%load_ext skip_kernel_extension

Reuse validation dataloader in accuracy testing for quantization. For that, it should be wrapped into the nncf.Dataset object and define a transformation function for getting only input tensors.

%%skip not $to_quantize.value

import nncf
from typing import Dict


def transform_fn(data_item:Dict):
    """
    Quantization transform function. Extracts and preprocess input data from dataloader item for quantization.
    Parameters:
       data_item: Dict with data item produced by DataLoader during iteration
    Returns:
        input_tensor: Input data for quantization
    """
    input_tensor = pose_validator.preprocess(data_item)['img'].numpy()
    return input_tensor


quantization_dataset = nncf.Dataset(pose_data_loader, transform_fn)

INFO:nncf:NNCF initialized successfully. Supported frameworks detected: torch, onnx, openvino

The nncf.quantize function provides an interface for model quantization. It requires an instance of the OpenVINO Model and quantization dataset. Optionally, some additional parameters for the configuration quantization process (number of samples for quantization, preset, ignored scope, etc.) can be provided. YOLOv8 model contains non-ReLU activation functions, which require asymmetric quantization of activations. To achieve a better result, we will use a mixed quantization preset. It provides symmetric quantization of weights and asymmetric quantization of activations. For more accurate results, we should keep the operation in the postprocessing subgraph in floating point precision, using the ignored_scope parameter.

Note: Model post-training quantization is time-consuming process. Be patient, it can take several minutes depending on your hardware.

%%skip not $to_quantize.value

ignored_scope = nncf.IgnoredScope(  # post-processing
    subgraphs=[
        nncf.Subgraph(inputs=['__module.model.22/aten::cat/Concat',
                              '__module.model.22/aten::cat/Concat_1',
                              '__module.model.22/aten::cat/Concat_2',
                             '__module.model.22/aten::cat/Concat_7'],
                      outputs=['__module.model.22/aten::cat/Concat_9'])
    ]
)

# Detection model
quantized_pose_model = nncf.quantize(
    pose_ov_model,
    quantization_dataset,
    preset=nncf.QuantizationPreset.MIXED,
    ignored_scope=ignored_scope
)

INFO:nncf:116 ignored nodes were found by subgraphs in the NNCFGraph
INFO:nncf:Not adding activation input quantizer for operation: 134 __module.model.22/aten::cat/Concat
INFO:nncf:Not adding activation input quantizer for operation: 142 __module.model.22/aten::view/Reshape_3
INFO:nncf:Not adding activation input quantizer for operation: 267 __module.model.22/aten::cat/Concat_1
INFO:nncf:Not adding activation input quantizer for operation: 279 __module.model.22/aten::view/Reshape_4
INFO:nncf:Not adding activation input quantizer for operation: 334 __module.model.22/aten::cat/Concat_2
INFO:nncf:Not adding activation input quantizer for operation: 337 __module.model.22/aten::view/Reshape_5
INFO:nncf:Not adding activation input quantizer for operation: 143 __module.model.22/aten::cat/Concat_7
INFO:nncf:Not adding activation input quantizer for operation: 154 __module.model.22/aten::view/Reshape_9
INFO:nncf:Not adding activation input quantizer for operation: 166 __module.model.22/aten::slice/Slice_2
INFO:nncf:Not adding activation input quantizer for operation: 167 __module.model.22/aten::slice/Slice_5
INFO:nncf:Not adding activation input quantizer for operation: 182 __module.model.22/aten::mul/Multiply_4
199 __module.model.22/aten::add/Add_8

INFO:nncf:Not adding activation input quantizer for operation: 183 __module.model.22/aten::sigmoid/Sigmoid_1
INFO:nncf:Not adding activation input quantizer for operation: 213 __module.model.22/aten::mul/Multiply_5
INFO:nncf:Not adding activation input quantizer for operation: 200 __module.model.22/aten::cat/Concat_8
INFO:nncf:Not adding activation input quantizer for operation: 214 __module.model.22/aten::view/Reshape_10
INFO:nncf:Not adding activation input quantizer for operation: 153 __module.model.22/aten::cat/Concat_4
INFO:nncf:Not adding activation input quantizer for operation: 165 __module.model.22/prim::ListUnpack
INFO:nncf:Not adding activation input quantizer for operation: 180 __module.model.22.dfl/aten::view/Reshape
INFO:nncf:Not adding activation input quantizer for operation: 181 __module.model.22/aten::sigmoid/Sigmoid
INFO:nncf:Not adding activation input quantizer for operation: 197 __module.model.22.dfl/aten::transpose/Transpose
INFO:nncf:Not adding activation input quantizer for operation: 211 __module.model.22.dfl/aten::softmax/Softmax
INFO:nncf:Not adding activation input quantizer for operation: 224 __module.model.22.dfl.conv/aten::_convolution/Convolution
INFO:nncf:Not adding activation input quantizer for operation: 232 __module.model.22.dfl/aten::view/Reshape_1
INFO:nncf:Not adding activation input quantizer for operation: 242 __module.model.22/prim::ListUnpack/VariadicSplit
INFO:nncf:Not adding activation input quantizer for operation: 251 __module.model.22/aten::sub/Subtract
INFO:nncf:Not adding activation input quantizer for operation: 252 __module.model.22/aten::add/Add_6
INFO:nncf:Not adding activation input quantizer for operation: 262 __module.model.22/aten::add/Add_7
273 __module.model.22/aten::div/Divide

INFO:nncf:Not adding activation input quantizer for operation: 263 __module.model.22/aten::sub/Subtract_1
INFO:nncf:Not adding activation input quantizer for operation: 274 __module.model.22/aten::cat/Concat_5
INFO:nncf:Not adding activation input quantizer for operation: 239 __module.model.22/aten::mul/Multiply_3
INFO:nncf:Not adding activation input quantizer for operation: 198 __module.model.22/aten::cat/Concat_9

Output()

Output()

%%skip not $to_quantize.value

print(f"Quantized keypoint detection model will be saved to {int8_model_pose_path}")
ov.save_model(quantized_pose_model, str(int8_model_pose_path))

Quantized keypoint detection model will be saved to models/yolov8n-pose_openvino_int8_model/yolov8n-pose.xml

Validate Quantized model inference

nncf.quantize returns the OpenVINO Model class instance, which is suitable for loading on a device for making predictions. INT8 model input data and output result formats have no difference from the floating point model representation. Therefore, we can reuse the same detect function defined above for getting the INT8 model result on the image.

%%skip not $to_quantize.value

device

%%skip not $to_quantize.value

ov_config = {}
if device.value != "CPU":
    quantized_pose_model.reshape({0: [1, 3, 640, 640]})
if "GPU" in device.value or ("AUTO" in device.value and "GPU" in core.available_devices):
    ov_config = {"GPU_DISABLE_WINOGRAD_CONVOLUTION": "YES"}
quantized_pose_compiled_model = core.compile_model(quantized_pose_model, device.value, ov_config)

def infer(*args):
    result = quantized_pose_compiled_model(args)
    return torch.from_numpy(result[0])

pose_model.predictor.inference = infer

res = pose_model(IMAGE_PATH)
display(Image.fromarray(res[0].plot()[:, :, ::-1]))

image 1/1 /home/akash/intel/openvino_notebooks/notebooks/yolov8-optimization/data/intel_rnb.jpg: 640x640 2 persons, 11.1ms
Speed: 3.8ms preprocess, 11.1ms inference, 1.4ms postprocess per image at shape (1, 3, 640, 640)

Compare the Original and Quantized Models

Compare performance of the Original and Quantized Models

Finally, use the OpenVINO Benchmark Tool to measure the inference performance of the FP32 and INT8 models.

Note: For more accurate performance, it is recommended to run benchmark_app in a terminal/command prompt after closing other applications. Run benchmark_app -m <model_path> -d CPU -shape "<input_shape>" to benchmark async inference on CPU on specific input data shape for one minute. Change CPU to GPU to benchmark on GPU. Run benchmark_app --help to see an overview of all command-line options.

%%skip not $to_quantize.value

device

if int8_model_pose_path.exists():
    # Inference FP32 model (OpenVINO IR)
    !benchmark_app -m $pose_model_path -d $device.value -api async -shape "[1,3,640,640]"

[Step 1/11] Parsing and validating input arguments
[ INFO ] Parsing input parameters
[Step 2/11] Loading OpenVINO Runtime
[ WARNING ] Default duration 120 seconds is used for unknown device AUTO
[ INFO ] OpenVINO:
[ INFO ] Build ................................. 2024.3.0-16041-1e3b88e4e3f-releases/2024/3
[ INFO ]
[ INFO ] Device info:
[ INFO ] AUTO
[ INFO ] Build ................................. 2024.3.0-16041-1e3b88e4e3f-releases/2024/3
[ INFO ]
[ INFO ]
[Step 3/11] Setting device configuration
[ WARNING ] Performance hint was not explicitly specified in command line. Device(AUTO) performance hint will be set to PerformanceMode.THROUGHPUT.
[Step 4/11] Reading model files
[ INFO ] Loading model files
[ INFO ] Read model took 30.04 ms
[ INFO ] Original model I/O parameters:
[ INFO ] Model inputs:
[ INFO ]     x (node: x) : f32 / [...] / [?,3,?,?]
[ INFO ] Model outputs:
[ INFO ]     *NO_NAME* (node: __module.model.22/aten::cat/Concat_9) : f32 / [...] / [?,56,21..]
[Step 5/11] Resizing model to match image sizes and given batch
[ INFO ] Model batch size: 1
[ INFO ] Reshaping model: 'x': [1,3,640,640]
[ INFO ] Reshape model took 8.90 ms
[Step 6/11] Configuring input of the model
[ INFO ] Model inputs:
[ INFO ]     x (node: x) : u8 / [N,C,H,W] / [1,3,640,640]
[ INFO ] Model outputs:
[ INFO ]     *NO_NAME* (node: __module.model.22/aten::cat/Concat_9) : f32 / [...] / [1,56,8400]
[Step 7/11] Loading the model to the device
[ INFO ] Compile model took 366.94 ms
[Step 8/11] Querying optimal runtime parameters
[ INFO ] Model:
[ INFO ]   NETWORK_NAME: Model0
[ INFO ]   EXECUTION_DEVICES: ['CPU']
[ INFO ]   PERFORMANCE_HINT: PerformanceMode.THROUGHPUT
[ INFO ]   OPTIMAL_NUMBER_OF_INFER_REQUESTS: 12
[ INFO ]   MULTI_DEVICE_PRIORITIES: CPU
[ INFO ]   CPU:
[ INFO ]     AFFINITY: Affinity.CORE
[ INFO ]     CPU_DENORMALS_OPTIMIZATION: False
[ INFO ]     CPU_SPARSE_WEIGHTS_DECOMPRESSION_RATE: 1.0
[ INFO ]     DYNAMIC_QUANTIZATION_GROUP_SIZE: 32
[ INFO ]     ENABLE_CPU_PINNING: True
[ INFO ]     ENABLE_HYPER_THREADING: True
[ INFO ]     EXECUTION_DEVICES: ['CPU']
[ INFO ]     EXECUTION_MODE_HINT: ExecutionMode.PERFORMANCE
[ INFO ]     INFERENCE_NUM_THREADS: 36
[ INFO ]     INFERENCE_PRECISION_HINT: <Type: 'float32'>
[ INFO ]     KV_CACHE_PRECISION: <Type: 'float16'>
[ INFO ]     LOG_LEVEL: Level.NO
[ INFO ]     MODEL_DISTRIBUTION_POLICY: set()
[ INFO ]     NETWORK_NAME: Model0
[ INFO ]     NUM_STREAMS: 12
[ INFO ]     OPTIMAL_NUMBER_OF_INFER_REQUESTS: 12
[ INFO ]     PERFORMANCE_HINT: THROUGHPUT
[ INFO ]     PERFORMANCE_HINT_NUM_REQUESTS: 0
[ INFO ]     PERF_COUNT: NO
[ INFO ]     SCHEDULING_CORE_TYPE: SchedulingCoreType.ANY_CORE
[ INFO ]   MODEL_PRIORITY: Priority.MEDIUM
[ INFO ]   LOADED_FROM_CACHE: False
[ INFO ]   PERF_COUNT: False
[Step 9/11] Creating infer requests and preparing input tensors
[ WARNING ] No input files were given for input 'x'!. This input will be filled with random values!
[ INFO ] Fill input 'x' with random values
[Step 10/11] Measuring performance (Start inference asynchronously, 12 inference requests, limits: 120000 ms duration)
[ INFO ] Benchmarking in inference only mode (inputs filling are not included in measurement loop).
[ INFO ] First inference took 41.78 ms
[Step 11/11] Dumping statistics report
[ INFO ] Execution Devices:['CPU']
[ INFO ] Count:            16740 iterations
[ INFO ] Duration:         120117.71 ms
[ INFO ] Latency:
[ INFO ]    Median:        85.71 ms
[ INFO ]    Average:       85.92 ms
[ INFO ]    Min:           64.95 ms
[ INFO ]    Max:           115.78 ms
[ INFO ] Throughput:   139.36 FPS

if int8_model_pose_path.exists():
    # Inference INT8 model (OpenVINO IR)
    !benchmark_app -m $int8_model_pose_path -d $device.value -api async -shape "[1,3,640,640]" -t 15

[Step 1/11] Parsing and validating input arguments
[ INFO ] Parsing input parameters
[Step 2/11] Loading OpenVINO Runtime
[ INFO ] OpenVINO:
[ INFO ] Build ................................. 2024.3.0-16041-1e3b88e4e3f-releases/2024/3
[ INFO ]
[ INFO ] Device info:
[ INFO ] AUTO
[ INFO ] Build ................................. 2024.3.0-16041-1e3b88e4e3f-releases/2024/3
[ INFO ]
[ INFO ]
[Step 3/11] Setting device configuration
[ WARNING ] Performance hint was not explicitly specified in command line. Device(AUTO) performance hint will be set to PerformanceMode.THROUGHPUT.
[Step 4/11] Reading model files
[ INFO ] Loading model files
[ INFO ] Read model took 20.82 ms
[ INFO ] Original model I/O parameters:
[ INFO ] Model inputs:
[ INFO ]     x (node: x) : f32 / [...] / [1,3,?,?]
[ INFO ] Model outputs:
[ INFO ]     *NO_NAME* (node: __module.model.22/aten::cat/Concat_9) : f32 / [...] / [1,56,21..]
[Step 5/11] Resizing model to match image sizes and given batch
[ INFO ] Model batch size: 1
[ INFO ] Reshaping model: 'x': [1,3,640,640]
[ INFO ] Reshape model took 11.11 ms
[Step 6/11] Configuring input of the model
[ INFO ] Model inputs:
[ INFO ]     x (node: x) : u8 / [N,C,H,W] / [1,3,640,640]
[ INFO ] Model outputs:
[ INFO ]     *NO_NAME* (node: __module.model.22/aten::cat/Concat_9) : f32 / [...] / [1,56,8400]
[Step 7/11] Loading the model to the device
[ INFO ] Compile model took 567.45 ms
[Step 8/11] Querying optimal runtime parameters
[ INFO ] Model:
[ INFO ]   NETWORK_NAME: Model0
[ INFO ]   EXECUTION_DEVICES: ['CPU']
[ INFO ]   PERFORMANCE_HINT: PerformanceMode.THROUGHPUT
[ INFO ]   OPTIMAL_NUMBER_OF_INFER_REQUESTS: 18
[ INFO ]   MULTI_DEVICE_PRIORITIES: CPU
[ INFO ]   CPU:
[ INFO ]     AFFINITY: Affinity.CORE
[ INFO ]     CPU_DENORMALS_OPTIMIZATION: False
[ INFO ]     CPU_SPARSE_WEIGHTS_DECOMPRESSION_RATE: 1.0
[ INFO ]     DYNAMIC_QUANTIZATION_GROUP_SIZE: 32
[ INFO ]     ENABLE_CPU_PINNING: True
[ INFO ]     ENABLE_HYPER_THREADING: True
[ INFO ]     EXECUTION_DEVICES: ['CPU']
[ INFO ]     EXECUTION_MODE_HINT: ExecutionMode.PERFORMANCE
[ INFO ]     INFERENCE_NUM_THREADS: 36
[ INFO ]     INFERENCE_PRECISION_HINT: <Type: 'float32'>
[ INFO ]     KV_CACHE_PRECISION: <Type: 'float16'>
[ INFO ]     LOG_LEVEL: Level.NO
[ INFO ]     MODEL_DISTRIBUTION_POLICY: set()
[ INFO ]     NETWORK_NAME: Model0
[ INFO ]     NUM_STREAMS: 18
[ INFO ]     OPTIMAL_NUMBER_OF_INFER_REQUESTS: 18
[ INFO ]     PERFORMANCE_HINT: THROUGHPUT
[ INFO ]     PERFORMANCE_HINT_NUM_REQUESTS: 0
[ INFO ]     PERF_COUNT: NO
[ INFO ]     SCHEDULING_CORE_TYPE: SchedulingCoreType.ANY_CORE
[ INFO ]   MODEL_PRIORITY: Priority.MEDIUM
[ INFO ]   LOADED_FROM_CACHE: False
[ INFO ]   PERF_COUNT: False
[Step 9/11] Creating infer requests and preparing input tensors
[ WARNING ] No input files were given for input 'x'!. This input will be filled with random values!
[ INFO ] Fill input 'x' with random values
[Step 10/11] Measuring performance (Start inference asynchronously, 18 inference requests, limits: 15000 ms duration)
[ INFO ] Benchmarking in inference only mode (inputs filling are not included in measurement loop).
[ INFO ] First inference took 30.80 ms
[Step 11/11] Dumping statistics report
[ INFO ] Execution Devices:['CPU']
[ INFO ] Count:            5688 iterations
[ INFO ] Duration:         15086.99 ms
[ INFO ] Latency:
[ INFO ]    Median:        46.98 ms
[ INFO ]    Average:       47.54 ms
[ INFO ]    Min:           33.22 ms
[ INFO ]    Max:           66.20 ms
[ INFO ] Throughput:   377.01 FPS

Compare accuracy of the Original and Quantized Models

As we can see, there is no significant difference between INT8 and float model result in a single image test. To understand how quantization influences model prediction precision, we can compare model accuracy on a dataset.

%%skip not $to_quantize.value

int8_pose_stats = test(quantized_pose_model, core, pose_data_loader, pose_validator, num_samples=NUM_TEST_SAMPLES)

0%|          | 0/300 [00:00<?, ?it/s]

%%skip not $to_quantize.value

print("FP32 model accuracy")
print_stats(fp_pose_stats, pose_validator.seen, pose_validator.nt_per_class.sum())

print("INT8 model accuracy")
print_stats(int8_pose_stats, pose_validator.seen, pose_validator.nt_per_class.sum())

FP32 model accuracy
Boxes:
    Best mean average:
               Class      Images      Labels   Precision      Recall      mAP@.5  mAP@.5:.95
                 all           8          21           1       0.899       0.955       0.736
INT8 model accuracy
Boxes:
    Best mean average:
               Class      Images      Labels   Precision      Recall      mAP@.5  mAP@.5:.95
                 all           8          21       0.886       0.952        0.98       0.723

Great! Looks like accuracy was changed, but not significantly and it meets passing criteria.

Other ways to optimize model

The performance could be also improved by another OpenVINO method such as async inference pipeline or preprocessing API.

Async Inference pipeline help to utilize the device more optimal. The key advantage of the Async API is that when a device is busy with inference, the application can perform other tasks in parallel (for example, populating inputs or scheduling other requests) rather than wait for the current inference to complete first. To understand how to perform async inference using openvino, refer to Async API tutorial

Preprocessing API enables making preprocessing a part of the model reducing application code and dependency on additional image processing libraries. The main advantage of Preprocessing API is that preprocessing steps will be integrated into the execution graph and will be performed on a selected device (CPU/GPU etc.) rather than always being executed on CPU as part of an application. This will also improve selected device utilization. For more information, refer to the overview of Preprocessing API tutorial. To see, how it could be used with YOLOV8 object detection model , please, see Convert and Optimize YOLOv8 real-time object detection with OpenVINO tutorial

Live demo

The following code runs model inference on a video:

import collections
import time
from IPython import display
import cv2


def run_keypoint_detection(
    source=0,
    flip=False,
    use_popup=False,
    skip_first_frames=0,
    model=pose_model,
    device=device.value,
):
    player = None

    ov_config = {}
    if device != "CPU":
        model.reshape({0: [1, 3, 640, 640]})
    if "GPU" in device or ("AUTO" in device and "GPU" in core.available_devices):
        ov_config = {"GPU_DISABLE_WINOGRAD_CONVOLUTION": "YES"}
    compiled_model = core.compile_model(model, device, ov_config)

    def infer(*args):
        result = compiled_model(args)
        return torch.from_numpy(result[0])

    pose_model.predictor.inference = infer

    try:
        # Create a video player to play with target fps.
        player = VideoPlayer(source=source, flip=flip, fps=30, skip_first_frames=skip_first_frames)
        # Start capturing.
        player.start()
        if use_popup:
            title = "Press ESC to Exit"
            cv2.namedWindow(winname=title, flags=cv2.WINDOW_GUI_NORMAL | cv2.WINDOW_AUTOSIZE)

        processing_times = collections.deque()
        while True:
            # Grab the frame.
            frame = player.next()
            if frame is None:
                print("Source ended")
                break
            # If the frame is larger than full HD, reduce size to improve the performance.
            scale = 1280 / max(frame.shape)
            if scale < 1:
                frame = cv2.resize(
                    src=frame,
                    dsize=None,
                    fx=scale,
                    fy=scale,
                    interpolation=cv2.INTER_AREA,
                )
            # Get the results
            input_image = np.array(frame)

            start_time = time.time()

            detections = pose_model(input_image)
            stop_time = time.time()
            frame = detections[0].plot()

            processing_times.append(stop_time - start_time)
            # Use processing times from last 200 frames.
            if len(processing_times) > 200:
                processing_times.popleft()

            _, f_width = frame.shape[:2]
            # Mean processing time [ms].
            processing_time = np.mean(processing_times) * 1000
            fps = 1000 / processing_time
            cv2.putText(
                img=frame,
                text=f"Inference time: {processing_time:.1f}ms ({fps:.1f} FPS)",
                org=(20, 40),
                fontFace=cv2.FONT_HERSHEY_COMPLEX,
                fontScale=f_width / 1000,
                color=(0, 0, 255),
                thickness=1,
                lineType=cv2.LINE_AA,
            )
            # Use this workaround if there is flickering.
            if use_popup:
                cv2.imshow(winname=title, mat=frame)
                key = cv2.waitKey(1)
                # escape = 27
                if key == 27:
                    break
            else:
                # Encode numpy array to jpg.
                _, encoded_img = cv2.imencode(ext=".jpg", img=frame, params=[cv2.IMWRITE_JPEG_QUALITY, 100])
                # Create an IPython image.
                i = display.Image(data=encoded_img)
                # Display the image in this notebook.
                display.clear_output(wait=True)
                display.display(i)
    # ctrl-c
    except KeyboardInterrupt:
        print("Interrupted")
    # any different error
    except RuntimeError as e:
        print(e)
    finally:
        if player is not None:
            # Stop capturing.
            player.stop()
        if use_popup:
            cv2.destroyAllWindows()

Run Keypoint Detection on video

VIDEO_SOURCE = "https://storage.openvinotoolkit.org/repositories/openvino_notebooks/data/data/video/people.mp4"

device

Dropdown(description='Device:', index=1, options=('CPU', 'AUTO'), value='AUTO')

run_keypoint_detection(
    source=VIDEO_SOURCE,
    flip=True,
    use_popup=False,
    model=pose_ov_model,
    device=device.value,
)

Source ended