search

TNTWEN / OpenVINO-YOLOV4

This is implementation of YOLOv4,YOLOv4-relu,YOLOv4-tiny,YOLOv4-tiny-3l,Scaled-YOLOv4 and INT8 Quantization in OpenVINO2021.3
MIT License
239 stars 66 forks source link

OpenVINO YOLOV4 FPS Discussion and Sharing #5

Closed TNTWEN closed 3 years ago

TNTWEN commented 4 years ago

Welcome to share the test results of OpenVINO-YOLOV4 on more Intel devices!

Test code ：object_detection_demo_yolov3_async.py

Test video: https://drive.google.com/file/d/1CMekikIu0W9jyFaNp6CXTQQ4hv3MQak4/view?usp=sharing

Size:416*416

TNTWEN commented 4 years ago

AI-BoxX Gen.1

YOLOv4 FP32：

DEVICE: HDDL-R8
USER_SPECIFIED mode  1.3FPS
MIN_LATENCY mode   1.3FPS

DEVICE: CPU - Intel Core i3-8145UE CPU @ 2.20GHz × 4 
USER_SPECIFIED mode   0.7FPS
MIN_LATENCY mode      0.7FPS

DEVICE:GPU - Intel UHD Graphics 620 (WHL GT2)
USER_SPECIFIED mode   2.3FPS
MIN_LATENCY mode      2.3FPS

YOLOv4 FP16：

DEVICE: HDDL-R8
USER_SPECIFIED mode   1.3FPS
MIN_LATENCY mode      1.3FPS

DEVICE: CPU - Intel Core i3-8145UE CPU @ 2.20GHz × 4 
USER_SPECIFIED mode   0.6/0.7FPS
MIN_LATENCY mode      0.6/0.7FPS

DEVICE:GPU - Intel UHD Graphics 620 (WHL GT2)
USER_SPECIFIED mode   3.1FPS
MIN_LATENCY mode      3.1FPS

YOLOv4-tiny FP32：

DEVICE: HDDL-R8
USER_SPECIFIED mode   14.2FPS
MIN_LATENCY mode      14.3FPS

DEVICE: CPU - Intel Core i3-8145UE CPU @ 2.20GHz × 4 
USER_SPECIFIED mode   9.3FPS
MIN_LATENCY mode      9.4FPS

DEVICE:GPU - Intel UHD Graphics 620 (WHL GT2)
USER_SPECIFIED mode   17.7FPS
MIN_LATENCY mode      18.0FPS

YOLOv4-tiny FP16：

DEVICE: HDDL-R8
USER_SPECIFIED mode   14.2FPS
MIN_LATENCY mode      14.2FPS

DEVICE: CPU - Intel Core i3-8145UE CPU @ 2.20GHz × 4 
USER_SPECIFIED mode   9.3FPS
MIN_LATENCY mode      9.4FPS

DEVICE:GPU - Intel UHD Graphics 620 (WHL GT2)
USER_SPECIFIED mode   21.2FPS
MIN_LATENCY mode      21.5FPS
jedibobo commented 4 years ago

Did you participate in Intel Cup for embedded devices? I've found the same hardware.