Two graphics cards installed in one computer.
and I checked with nvidia-smi.
I ran this simple code on both graphics cards to see if it worked.
model = YOLO('yolov8x-pose.pt') def pose_detection(image_path): file_name, ext = os.path.splitext(image_path) image = cv2.imread(image_path) height, width, channels = image.shape results = model.predict(image, save=False,device=0,half=True, iou=0.5, conf=0.3) #results = model.predict(image, save=False,device=1,half=True, iou=0.5, conf=0.3) boxes = results[0].boxes.xyxy.cpu().numpy().astype(int) print(len(boxes)) image_path='./test12_13_0.jpg' for i in range(1, 1000): print(i) pose_detection(image_path) I set device=0 and device=1 respectively and ran the code.
But I don't understand why this result occurs.
0: 384x640 7 persons, 23.6ms Speed: 1.7ms preprocess, 23.6ms inference, 1.0ms postprocess per image at shape (1, 3, 384, 640) 7 0: 384x640 7 persons, 10.4ms Speed: 1.9ms preprocess, 10.4ms inference, 1.3ms postprocess per image at shape (1, 3, 384, 640) 7 There is also a difference in recognition time. (The one at the top is 6000 ada.)
Wasn't the performance of the RTX 6000 ADA better than the RTX A6000? Or is it wrong to measure with this code?
I would appreciate it if someone could let me know.
2Related questions 4 RTX 3070 compatibility with Pytorch 1 How is the NVIDIA RTX A6000 getting 300 watts via a single 8-pin? 0 Why isn't Blender using the GeForce RTX 2080 GPU and using only the Intel UHD Graphics 630? Related questions 4 RTX 3070 compatibility with Pytorch 1 How is the NVIDIA RTX A6000 getting 300 watts via a single 8-pin? 0 Why isn't Blender using the GeForce RTX 2080 GPU and using only the Intel UHD Graphics 630? 0 How to Install Cuda 10.1 with Tensorflow V.2.4 RTX 2070 Super Ubuntu 18.04 Load 1 more related questions Show fewer related questions
Reset to default
