File size: 4,007 Bytes
ac9dbb5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
4a46c88
ac9dbb5
4a46c88
ac9dbb5
 
4a46c88
 
 
 
 
 
 
 
 
 
 
ac9dbb5
 
 
 
 
 
 
 
 
4a46c88
 
 
 
 
 
 
 
 
 
 
69e7b1f
 
 
 
 
 
 
 
ac9dbb5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
807b82c
ac9dbb5
 
 
 
807b82c
 
 
 
 
 
 
 
 
 
 
 
 
 
ac9dbb5
807b82c
ac9dbb5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
#!/usr/bin/env bash
# SPDX-License-Identifier: MIT
# Copyright (C) Intel Corporation
#
# Export a YOLO26 detector for motion tracking to OpenVINO IR.
# The tracker itself (BoT-SORT / ByteTrack) runs at inference time via
# Ultralytics or DLStreamer gvatrack; no separate model export is needed.
# Usage: ./export_and_quantize.sh [MODEL_VARIANT] [PRECISION]
# Example: ./export_and_quantize.sh yolo26n FP16
#
# Supported precisions:
#   FP32  -- Full-precision floating-point weights
#   FP16  -- Half-precision floating-point weights (default)
#   INT8  -- Quantized 8-bit integer weights (requires NNCF)
#
# Precision / device compatibility:
#   | Precision | CPU | GPU | NPU |
#   |-----------|-----|-----|-----|
#   | FP32      | Yes | Yes | No  |
#   | FP16      | Yes | Yes | Yes |
#   | INT8      | Yes | Yes | Yes |

set -euo pipefail

MODEL_NAME="${1:-yolo26n}"
PRECISION="${2:-FP16}"
PRECISION="$(echo "${PRECISION}" | tr '[:lower:]' '[:upper:]')"

if [[ "${PRECISION}" != "FP32" && "${PRECISION}" != "FP16" && "${PRECISION}" != "INT8" ]]; then
    echo "ERROR: unsupported precision '${PRECISION}'. Choose FP32, FP16, or INT8." >&2
    exit 1
fi

echo "--- Installing dependencies ---"
if [[ "${PRECISION}" == "INT8" ]]; then
    pip install -qU openvino nncf ultralytics
else
    pip install -qU openvino ultralytics
fi

# Ask for approval before downloading models and sample files
echo ""
echo "This script will download:"
echo "  - Model weights and/or sample files"
echo ""
read -p "Continue with downloads? (yes/no): " APPROVAL
if [[ "${APPROVAL}" != "yes" ]]; then
    echo "Download cancelled by user."
    exit 0
fi
echo ""
echo "--- Downloading sample test video ---"
if [[ ! -f test_video.mp4 ]]; then
    wget -q -O test_video.mp4 \
        https://github.com/intel-iot-devkit/sample-videos/raw/master/people-detection.mp4
    echo "Downloaded: test_video.mp4"
else
    echo "Already present: test_video.mp4"
fi

# Ask for approval before downloading models and sample files
echo ""
echo "This script will download:"
echo "  - Model weights and/or sample files"
echo ""
read -p "Continue with downloads? (yes/no): " APPROVAL
if [[ "${APPROVAL}" != "yes" ]]; then
    echo "Download cancelled by user."
    exit 0
fi
echo ""
echo "--- Downloading sample test image ---"
if [[ ! -f test.jpg ]]; then
    wget -q -O test.jpg https://ultralytics.com/images/bus.jpg
    echo "Downloaded: test.jpg"
else
    echo "Already present: test.jpg"
fi

if [[ "${PRECISION}" == "FP32" ]]; then
    HALF_FLAG="False"
    EXPORT_LABEL="FP32"
else
    HALF_FLAG="True"
    EXPORT_LABEL="FP16"
fi

echo "--- Exporting ${MODEL_NAME} to OpenVINO IR (${EXPORT_LABEL}) ---"
python3 -c "
from ultralytics import YOLO

model = YOLO('${MODEL_NAME}.pt')
model.export(format='openvino', half=${HALF_FLAG}, dynamic=False, imgsz=640)
print('Export complete: ${MODEL_NAME}_openvino_model/')
"

if [[ "${PRECISION}" == "INT8" ]]; then
    echo "--- Quantizing to INT8 with NNCF ---"
    python3 -c "
import nncf
import openvino as ov
import numpy as np
import cv2

core = ov.Core()
model = core.read_model('${MODEL_NAME}_openvino_model/${MODEL_NAME}.xml')

# Extract frames from the sample video for calibration.
cap = cv2.VideoCapture('test_video.mp4')
frames = []
while len(frames) < 300:
    ret, frame = cap.read()
    if not ret:
        cap.set(cv2.CAP_PROP_POS_FRAMES, 0)
        continue
    img = cv2.resize(frame, (640, 640))
    img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB).astype(np.float32) / 255.0
    img = img.transpose(2, 0, 1)[np.newaxis, ...]
    frames.append(img)
cap.release()

def transform_fn(data_item):
    return frames[data_item % len(frames)]

calibration_dataset = nncf.Dataset(list(range(300)), transform_fn)

quantized = nncf.quantize(
    model,
    calibration_dataset,
    preset=nncf.QuantizationPreset.MIXED,
    subset_size=300,
)

ov.save_model(quantized, '${MODEL_NAME}_tracking_int8.xml')
print('Quantization complete: ${MODEL_NAME}_tracking_int8.xml')
"
fi
echo "--- Done ---"