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¶
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==8.1.42" "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
# 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/maleksandr/test_notebooks/update_ultralytics/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 https://github.com/ultralytics/assets/releases/download/v8.1.0/yolov8n-pose.pt to 'models/yolov8n-pose.pt'...
100%|██████████████████████████████████████████████████████████████████████████████| 6.51M/6.51M [00:01<00:00, 3.93MB/s]
image 1/1 /home/maleksandr/test_notebooks/update_ultralytics/openvino_notebooks/notebooks/yolov8-optimization/data/intel_rnb.jpg: 480x640 1 person, 45.3ms
Speed: 1.5ms preprocess, 45.3ms inference, 1.0ms 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 YOLOv8.1.42 🚀 Python-3.10.12 torch-2.2.2+cpu 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 2024.0.0-14509-34caeefd078-releases/2024/0...
OpenVINO: export success ✅ 1.7s, saved as 'models/yolov8n-pose_openvino_model/' (6.7 MB)
Export complete (3.0s)
Results saved to /home/maleksandr/test_notebooks/update_ultralytics/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
import ipywidgets as widgets
import openvino as ov
core = ov.Core()
device = widgets.Dropdown(
options=core.available_devices + ["AUTO"],
value="AUTO",
description="Device:",
disabled=False,
)
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
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/maleksandr/test_notebooks/update_ultralytics/openvino_notebooks/notebooks/yolov8-optimization/data/intel_rnb.jpg: 640x640 2 persons, 23.7ms
Speed: 5.0ms preprocess, 23.7ms inference, 2.2ms 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/maleksandr/test_notebooks/ultrali/datasets/val2017.zip: 0%| | 0.00/334k [00:00<?, ?B/s]
/home/maleksandr/test_notebooks/ultrali/datasets/coco8-pose.yaml: 0%| | 0.00/552 [00:00<?, ?B/s]
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/maleksandr/test_notebooks/ultrali/datasets/coco8-pose/labels/train.cache... 8 images, 0 backgrounds,
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:
Precisionis the degree of exactness of the model in identifying only relevant objects.Recallmeasures 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, wheretis the Intersection Over Union (IOU) threshold, degree of overlapping between ground truth and predicted objects. Therefore,mAP@.5indicates 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:
Create a Dataset for quantization.
Run
nncf.quantizefor getting an optimized model.Serialize OpenVINO IR model, using the
openvino.runtime.serializefunction.
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(
names=[
"__module.model.22.cv3.0.0.conv/aten::_convolution/Convolution", # in the post-processing subgraph
"__module.model.22.cv4.0.0.conv/aten::_convolution/Convolution",
"__module.model.16.conv/aten::_convolution/Convolution",
"__module.model.22.cv2.0.0.conv/aten::_convolution/Convolution",
"__module.model.6.cv1.conv/aten::_convolution/Convolution",
"__module.model.22.cv3.1.1.conv/aten::_convolution/Convolution",
"__module.model.21.cv2.conv/aten::_convolution/Convolution",
"__module.model.21.m.0.cv1.conv/aten::_convolution/Convolution",
"__module.model.22/aten::add/Add_6",
"__module.model.22/aten::sub/Subtract",
"__module.model.7.conv/aten::_convolution/Convolution",
"__module.model.12.cv1.conv/aten::_convolution/Convolution",
"__module.model.4.cv1.conv/aten::_convolution/Convolution",
"__module.model.22.cv2.2.1.conv/aten::_convolution/Convolution",
"__module.model.22.cv2.0.1.conv/aten::_convolution/Convolution",
"__module.model.22.cv4.2.1.conv/aten::_convolution/Convolution",
"__module.model.22.dfl.conv/aten::_convolution/Convolution",
"__module.model.22.cv3.2.2/aten::_convolution/Convolution",
"__module.model.22.cv3.0.2/aten::_convolution/Convolution",
"__module.model.15.cv1.conv/aten::_convolution/Convolution",
"__module.model.5.conv/aten::_convolution/Convolution",
"__module.model.0.conv/aten::_convolution/Convolution"
]
)
# Detection model
quantized_pose_model = nncf.quantize(
pose_ov_model,
quantization_dataset,
preset=nncf.QuantizationPreset.MIXED,
ignored_scope=ignored_scope
)
INFO:nncf:22 ignored nodes were found by name in the NNCFGraph
INFO:nncf:Not adding activation input quantizer for operation: 1 __module.model.0.conv/aten::_convolution/Convolution
2 __module.model.0.conv/aten::_convolution/Add
3 __module.model.22.cv4.2.1.act/aten::silu_/Swish
INFO:nncf:Not adding activation input quantizer for operation: 25 __module.model.4.cv1.conv/aten::_convolution/Convolution
26 __module.model.4.cv1.conv/aten::_convolution/Add
27 __module.model.22.cv4.2.1.act/aten::silu_/Swish_7
INFO:nncf:Not adding activation input quantizer for operation: 43 __module.model.5.conv/aten::_convolution/Convolution
47 __module.model.5.conv/aten::_convolution/Add
51 __module.model.22.cv4.2.1.act/aten::silu_/Swish_13
INFO:nncf:Not adding activation input quantizer for operation: 54 __module.model.6.cv1.conv/aten::_convolution/Convolution
56 __module.model.6.cv1.conv/aten::_convolution/Add
59 __module.model.22.cv4.2.1.act/aten::silu_/Swish_14
INFO:nncf:Not adding activation input quantizer for operation: 99 __module.model.7.conv/aten::_convolution/Convolution
112 __module.model.7.conv/aten::_convolution/Add
123 __module.model.22.cv4.2.1.act/aten::silu_/Swish_20
INFO:nncf:Not adding activation input quantizer for operation: 111 __module.model.12.cv1.conv/aten::_convolution/Convolution
122 __module.model.12.cv1.conv/aten::_convolution/Add
131 __module.model.22.cv4.2.1.act/aten::silu_/Swish_27
INFO:nncf:Not adding activation input quantizer for operation: 46 __module.model.15.cv1.conv/aten::_convolution/Convolution
50 __module.model.15.cv1.conv/aten::_convolution/Add
53 __module.model.22.cv4.2.1.act/aten::silu_/Swish_31
INFO:nncf:Not adding activation input quantizer for operation: 74 __module.model.16.conv/aten::_convolution/Convolution
83 __module.model.16.conv/aten::_convolution/Add
92 __module.model.22.cv4.2.1.act/aten::silu_/Swish_39
INFO:nncf:Not adding activation input quantizer for operation: 75 __module.model.22.cv2.0.0.conv/aten::_convolution/Convolution
84 __module.model.22.cv2.0.0.conv/aten::_convolution/Add
93 __module.model.22.cv4.2.1.act/aten::silu_/Swish_35
INFO:nncf:Not adding activation input quantizer for operation: 103 __module.model.22.cv2.0.1.conv/aten::_convolution/Convolution
116 __module.model.22.cv2.0.1.conv/aten::_convolution/Add
126 __module.model.22.cv4.2.1.act/aten::silu_/Swish_36
INFO:nncf:Not adding activation input quantizer for operation: 76 __module.model.22.cv3.0.0.conv/aten::_convolution/Convolution
85 __module.model.22.cv3.0.0.conv/aten::_convolution/Add
94 __module.model.22.cv4.2.1.act/aten::silu_/Swish_37
INFO:nncf:Not adding activation input quantizer for operation: 135 __module.model.22.cv3.0.2/aten::_convolution/Convolution
143 __module.model.22.cv3.0.2/aten::_convolution/Add
INFO:nncf:Not adding activation input quantizer for operation: 77 __module.model.22.cv4.0.0.conv/aten::_convolution/Convolution
86 __module.model.22.cv4.0.0.conv/aten::_convolution/Add
95 __module.model.22.cv4.2.1.act/aten::silu_/Swish_57
INFO:nncf:Not adding activation input quantizer for operation: 234 __module.model.22.cv3.1.1.conv/aten::_convolution/Convolution
247 __module.model.22.cv3.1.1.conv/aten::_convolution/Add
258 __module.model.22.cv4.2.1.act/aten::silu_/Swish_47
INFO:nncf:Not adding activation input quantizer for operation: 289 __module.model.21.m.0.cv1.conv/aten::_convolution/Convolution
296 __module.model.21.m.0.cv1.conv/aten::_convolution/Add
302 __module.model.22.cv4.2.1.act/aten::silu_/Swish_50
INFO:nncf:Not adding activation input quantizer for operation: 295 __module.model.21.cv2.conv/aten::_convolution/Convolution
301 __module.model.21.cv2.conv/aten::_convolution/Add
305 __module.model.22.cv4.2.1.act/aten::silu_/Swish_52
INFO:nncf:Not adding activation input quantizer for operation: 332 __module.model.22.cv2.2.1.conv/aten::_convolution/Convolution
340 __module.model.22.cv2.2.1.conv/aten::_convolution/Add
345 __module.model.22.cv4.2.1.act/aten::silu_/Swish_54
INFO:nncf:Not adding activation input quantizer for operation: 334 __module.model.22.cv4.2.1.conv/aten::_convolution/Convolution
342 __module.model.22.cv4.2.1.conv/aten::_convolution/Add
347 __module.model.22.cv4.2.1.act/aten::silu_/Swish_62
INFO:nncf:Not adding activation input quantizer for operation: 350 __module.model.22.cv3.2.2/aten::_convolution/Convolution
354 __module.model.22.cv3.2.2/aten::_convolution/Add
INFO:nncf:Not adding activation input quantizer for operation: 243 __module.model.22.dfl.conv/aten::_convolution/Convolution
INFO:nncf:Not adding activation input quantizer for operation: 263 __module.model.22/aten::sub/Subtract
INFO:nncf:Not adding activation input quantizer for operation: 264 __module.model.22/aten::add/Add_6
Output()
/home/maleksandr/test_notebooks/update_ultralytics/openvino_notebooks/notebooks/yolov8-optimization/venv/lib/python3.10/site-packages/nncf/experimental/tensor/tensor.py:84: RuntimeWarning: invalid value encountered in multiply
return Tensor(self.data * unwrap_tensor_data(other))
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/maleksandr/test_notebooks/update_ultralytics/openvino_notebooks/notebooks/yolov8-optimization/data/intel_rnb.jpg: 640x640 2 persons, 20.0ms
Speed: 3.5ms preprocess, 20.0ms inference, 1.1ms 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_appin a terminal/command prompt after closing other applications. Runbenchmark_app -m <model_path> -d CPU -shape "<input_shape>"to benchmark async inference on CPU on specific input data shape for one minute. ChangeCPUtoGPUto benchmark on GPU. Runbenchmark_app --helpto 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 ................................. 2024.0.0-14509-34caeefd078-releases/2024/0 [ INFO ] [ INFO ] Device info: [ INFO ] AUTO [ INFO ] Build ................................. 2024.0.0-14509-34caeefd078-releases/2024/0 [ 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 16.24 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 / [...] / [?,56,16..] [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 9.32 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 313.69 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: 0 [ 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 ] 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 [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 42.40 ms [Step 11/11] Dumping statistics report [ INFO ] Execution Devices:['CPU'] [ INFO ] Count: 21612 iterations [ INFO ] Duration: 120076.97 ms [ INFO ] Latency: [ INFO ] Median: 66.18 ms [ INFO ] Average: 66.51 ms [ INFO ] Min: 32.58 ms [ INFO ] Max: 123.83 ms [ INFO ] Throughput: 179.98 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 ................................. 2024.0.0-14509-34caeefd078-releases/2024/0 [ INFO ] [ INFO ] Device info: [ INFO ] AUTO [ INFO ] Build ................................. 2024.0.0-14509-34caeefd078-releases/2024/0 [ 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 23.83 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 13.45 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 540.15 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: 0 [ 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 ] 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 [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: 15000 ms duration) [ INFO ] Benchmarking in inference only mode (inputs filling are not included in measurement loop). [ INFO ] First inference took 34.70 ms [Step 11/11] Dumping statistics report [ INFO ] Execution Devices:['CPU'] [ INFO ] Count: 4248 iterations [ INFO ] Duration: 15064.81 ms [ INFO ] Latency: [ INFO ] Median: 42.11 ms [ INFO ] Average: 42.37 ms [ INFO ] Min: 24.13 ms [ INFO ] Max: 69.95 ms [ INFO ] Throughput: 281.98 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 0.952 0.938 0.953 0.638
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