Runtime initialized with Backend::ORT in Device:: CPU 和 GPU 有问题

[monkeycc]

环境

FastDeploy版本： fastdeploy-gpu-python 0.6.0 pip 系统平台: Windows x64(Windows11) 硬件： cuda11.6 cudnn8.4 编译语言： Python(3.7） 问题描述

问题描述

import cv2
import fastdeploy.vision as vision

model = vision.detection.PPYOLOE("ppyoloe_crn_l_300e_coco/model.pdmodel",
                                 "ppyoloe_crn_l_300e_coco/model.pdiparams",
                                 "ppyoloe_crn_l_300e_coco/infer_cfg.yml")
im = cv2.imread("000000014439.jpg")
result = model.predict(im.copy())
print(result)

vis_im = vision.vis_detection(im, result, score_threshold=0.5)
cv2.imwrite("vis_image.jpg", vis_im)

显示 Runtime initialized with Backend::ORT in Device:: CPU

---

import cv2
import fastdeploy as fd

def build_option():
    option = fd.RuntimeOption()

    # if args.device.lower() == "gpu":
    option.use_gpu()

    # if args.use_trt:
    #     option.use_trt_backend()
    return option

runtime_option = build_option()
model = vision.detection.PPYOLOE("ppyoloe_crn_l_300e_coco/model.pdmodel",
                                 "ppyoloe_crn_l_300e_coco/model.pdiparams",
                                 "ppyoloe_crn_l_300e_coco/infer_cfg.yml", runtime_option=runtime_option)
im = cv2.imread("000000014439.jpg")
result = model.predict(im.copy())
print(result)

vis_im = vision.vis_detection(im, result, score_threshold=0.5)
cv2.imwrite("vis_image.jpg", vis_im)

显示 Runtime initialized with Backend::ORT in Device:: GPU

---

发现 用 fastdeploy.vision as vision 只会显示 CPU 需要用import fastdeploy as fd 才会显示GPU

这两个 有什么区别的

[jiangjiajun]

因为后面那段代码构造PPYOLOE时传入了runtime_option，设置了要在GPU上跑，默认是CPU。 与怎么import没有关系。

[DefTruth]

option.use_gpu()

这句开启了GPU推理

[jiangjiajun]

此ISSUE由于一年未更新，将会关闭处理，如有需要，可再次更新打开。