Ultralytics
modes/predict/ #7932
giscus
bot
Replies: 111 comments 248 replies
-
|
thx, its amazing |
Beta Was this translation helpful? Give feedback.
-
|
@soapboss you're welcome! If you have any more questions or need further assistance, feel free to ask. Happy to help! 😊🚀 |
Beta Was this translation helpful? Give feedback.
-
|
Can we set different confidences for different classes? It would be a nice addition. |
Beta Was this translation helpful? Give feedback.
-
|
@jwmetrifork currently, YOLOv8 does not support setting different confidence thresholds for different classes directly through the model's configuration or command-line arguments. The confidence threshold is a global setting that applies to all classes equally. However, you can implement custom post-processing logic in Python after running predictions with the model. Here's a basic example of how you might apply different confidence thresholds for different classes after obtaining the results: from ultralytics import YOLO
# Load a pretrained YOLOv8n model
model = YOLO('yolov8n.pt')
# Run inference on an image
results = model('bus.jpg')
# Define custom confidence thresholds for each class (example)
custom_conf_thresholds = {
0: 0.5, # Class ID 0 with 0.5 confidence
1: 0.3, # Class ID 1 with 0.3 confidence
# Add more classes and thresholds as needed
}
# Filter results based on custom confidence thresholds
filtered_boxes = []
for box in results[0].boxes:
class_id = int(box.cls)
if box.conf >= custom_conf_thresholds.get(class_id, 0.25): # Default threshold if class ID not in dict
filtered_boxes.append(box)
# Do something with the filtered results
# ...In this example, If you need this functionality often, consider suggesting it as a feature request on the YOLOv8 GitHub repository. The maintainers and community can discuss the feasibility and potential implementation of such a feature. 🚀 |
Beta Was this translation helpful? Give feedback.
-
|
Thanks for answering. Yeah, I do something similar also. I will look into suggesting it as a feature |
Beta Was this translation helpful? Give feedback.
-
|
@jwmetrifork you're welcome! Custom post-processing like you're doing is a great way to tailor the model's output to your specific needs. If you feel that different confidence thresholds for different classes would be a valuable addition to YOLOv8, definitely consider suggesting it as a feature request on the GitHub repository. Your input could help improve the tool for everyone. If you need any assistance with the feature request or have further questions, feel free to reach out. Happy coding! 😊👨💻 |
Beta Was this translation helpful? Give feedback.
-
|
Can you please share more details on the I understand that by default we receive 160x160 masks in segment models, but what does setting the above True do? |
Beta Was this translation helpful? Give feedback.
-
|
@adbcode certainly! The When you set the However, it's important to note that using Here's an example of how to use the from ultralytics import YOLO
# Load a pretrained segmentation model
model = YOLO('yolov8n-seg.pt')
# Run inference with high-resolution masks
results = model.predict('image.jpg', retina_masks=True)In this example, the segmentation masks returned in the |
Beta Was this translation helpful? Give feedback.
-
|
Hi @pderrenger ! Thanks for a quick response. I noticed that the trained model works on images with larger resolution than the So to confirm, will the masks be created in Regardless, I now understand that we can have more precise masks with this property. |
Beta Was this translation helpful? Give feedback.
-
|
Hi there! 👋 Yes, during prediction, if the Hope that clarifies it! If you have any more questions, feel free to ask. 😊 |
Beta Was this translation helpful? Give feedback.
-
|
Hi, I am trying to do a simple yolo model and every time I try predicting it doesn't detect anything. Here is what I am doing. I am. downloading this dataset and then trying to train a model on it and then once its trained I am trying to predict with it. I pasted all of my code so that you can replicate. https://universe.roboflow.com/damage-4yhkc/damaged-bchyj/dataset/5/downloadfrom roboflow import Roboflow model = YOLO('yolov8n-seg.pt') # build from YAML and transfer weights #Predict |
Beta Was this translation helpful? Give feedback.
-
|
@goshua13 hello! It seems like you've put in a lot of effort into getting your YOLO model to predict correctly. If it's working in Google Colab but not in your local setup, there might be an issue with the environment or the way the model is being loaded for prediction. Here are a few things you can check:
If you're still facing issues, consider sharing the error messages or the behavior you're observing. This can provide more context for troubleshooting. Also, you can refer to the Ultralytics YOLOv8 documentation for more detailed guidance on predictions. Keep experimenting, and don't hesitate to reach out to the community for more specific advice! |
Beta Was this translation helpful? Give feedback.
-
|
Awesome thank you, I am trying to save each image when I predict, and right now I am just using the save_crop=True, and its working great the only thing is that It doesn't save the metadata of the original image. I need the metadata from the original image to be saved for each cropped image. Is there a way to do this? |
Beta Was this translation helpful? Give feedback.
-
|
Hey there! 👋 Great to hear that Here's a quick Python snippet using from PIL import Image
# Load the original image and its metadata
original = Image.open('path/to/original/image.jpg')
metadata = original.info
# Load the cropped image
cropped = Image.open('path/to/cropped/image.jpg')
# Save the cropped image with the original metadata
cropped.save('path/to/save/cropped_with_metadata.jpg', **metadata)This will copy the metadata from the original image to the cropped one. Just loop through your cropped images and apply this snippet to each. Hope this helps! 😊 Let me know if you need any more assistance. Happy coding! |
Beta Was this translation helpful? Give feedback.
-
|
great! thank you! |
Beta Was this translation helpful? Give feedback.
-
|
Hey, thanks for reaching out! 🌟 For saving cropped images with the original metadata, you might consider a post-processing step where you copy the metadata from the original images to the cropped ones after prediction. This isn't handled by default in YOLOv8, but you can use libraries like Here's a quick example using import piexif
# Load metadata from the original image
original_metadata = piexif.load('original_image.jpg')
# Copy metadata to the cropped image
piexif.insert(piexif.dump(original_metadata), 'cropped_image.jpg')Just loop this for each cropped image. Hope this helps and happy coding! 😊👨💻 |
Beta Was this translation helpful? Give feedback.
-
|
Hello if i want to know these pixel coordinates represent which object, What should i do? For example this is my code And this code give to me the pixel coordinates of objects so how can i learn which coordinate represents which object? |
Beta Was this translation helpful? Give feedback.
-
|
Hey there! 👋 To find out which object corresponds to a particular set of pixel coordinates, you can check the from ultralytics import YOLO
# Load a pretrained YOLOv8n-seg Segment model
model = YOLO('yolov8n-seg.pt')
# Run inference on an image
results = model('bus.jpg') # results list
# View results
for r in results:
for i, mask in enumerate(r.masks.xy):
class_id = r.boxes.cls[i].item() # Get class ID for each mask
class_name = r.names[class_id] # Convert class ID to class name
print(f"Object: {class_name}, Mask Coordinates: {mask}")This will print out the class name for each set of mask coordinates. Hope this helps! 😊 |
Beta Was this translation helpful? Give feedback.
-
|
With that code : it print that results: |
Beta Was this translation helpful? Give feedback.
-
|
Hey there! 👋 It looks like you're trying to access the Detection models like from ultralytics import YOLO
# Load a classification model
model = YOLO('yolov8n-cls.pt') # pretrained YOLOv8n classification model
# Run inference on images
results = model(['frame_1147.jpg', 'frame_2527.jpg'])
# Process results list
for result in results:
probs = result.probs # Probs object for classification outputs
print(probs.top5) # Print top-5 class probabilitiesHope this clears things up! 😊 Let me know if you have any other questions. |
Beta Was this translation helpful? Give feedback.
-
|
how can we count total number of objects in the set of images after training the model? |
Beta Was this translation helpful? Give feedback.
-
|
Hey there! 👋 To count the total number of objects across a set of images after training your model, you can use the from ultralytics import YOLO
# Load your trained model
model = YOLO('path/to/your/trained_model.pt')
# List of images to run inference on
images = ['image1.jpg', 'image2.jpg', ...]
# Run inference
results = model(images)
# Count total objects
total_objects = sum(len(result.boxes) for result in results)
print(f'Total objects detected: {total_objects}')This will give you the total count of objects detected across all your images. Happy counting! 😊 |
Beta Was this translation helpful? Give feedback.
-
|
Thank you so much for your response. I appreciate it.
I want to know can we use this code for directory because I have 760 images
in the directory I want count the total predicted boxes all in once.
Thanks
…On Tue, Feb 13, 2024, 9:55 PM Glenn Jocher ***@***.***> wrote:
Hey there! 👋 To count the total number of objects across a set of images
after training your model, you can use the predict mode to process your
images and then sum up the detections. Here's a quick example using Python:
from ultralytics import YOLO
# Load your trained modelmodel = YOLO('path/to/your/trained_model.pt')
# List of images to run inference onimages = ['image1.jpg', 'image2.jpg', ...]
# Run inferenceresults = model(images)
# Count total objectstotal_objects = sum(len(result.boxes) for result in results)print(f'Total objects detected: {total_objects}')
This will give you the total count of objects detected across all your
images. Happy counting! 😊
—
Reply to this email directly, view it on GitHub
<#7932 (reply in thread)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/BAHRMJFIOOEU34WCWMTYWELYTNWFDAVCNFSM6AAAAABCTDRDM6VHI2DSMVQWIX3LMV43SRDJONRXK43TNFXW4Q3PNVWWK3TUHM4DINJUGA2TM>
.
You are receiving this because you commented.Message ID:
***@***.***
com>
|
Beta Was this translation helpful? Give feedback.
-
|
@mudassir659 absolutely, you can use YOLOv8 to process a directory with multiple images and count the total predicted boxes. Here's a simple way to do it using Python: from ultralytics import YOLO
# Load your trained model
model = YOLO('path/to/your/trained_model.pt')
# Run inference on the directory
results = model('path/to/your/directory', stream=True) # Use stream=True for efficiency
# Count total boxes
total_boxes = sum(len(result.boxes) for result in results)
print(f'Total predicted boxes: {total_boxes}')This code will iterate over all images in the specified directory, perform predictions, and tally up the total number of detected boxes. Let me know if you need any more help! 😊👍 |
Beta Was this translation helpful? Give feedback.
-
|
If i want to run predictions on several images at once, I shoumd use a tensor as |
Beta Was this translation helpful? Give feedback.
-
|
Hey there! 👋 For running predictions on multiple images of varying sizes and aspect ratios, you can pad each image to the same size or use the import torch
from PIL import Image
# Load your images
images = [Image.open(path) for path in ['image1.jpg', 'image2.jpg']]
# Pad images to the same size
max_size = tuple(max(sizes) for sizes in zip(*[img.size for img in images]))
padded_images = [torch.nn.functional.pad(torch.from_numpy(np.array(img)),
(0, max_size[0] - img.size[0], 0, max_size[1] - img.size[1]),
mode='constant', value=0) for img in images]
# Stack images into a single tensor
input_tensor = torch.stack(padded_images).float() / 255.0 # Normalize to [0, 1]Now Let me know if you need further assistance! |
Beta Was this translation helpful? Give feedback.
-
|
If I want to predict on several images at once I should use a tensor with the format |
Beta Was this translation helpful? Give feedback.
-
|
@vonzo hey there! 👋 For predicting on a batch of images with different sizes and ratios using a tensor in Here's a quick example of how you might prepare your images: import torch
from PIL import Image
from torchvision.transforms import functional as F
# Load your images
images = [Image.open(path) for path in ['image1.jpg', 'image2.jpg', ...]]
# Resize or pad them to the same size
processed_images = [F.to_tensor(F.resize(img, (640, 640))) for img in images]
# Stack into a single tensor
batch_tensor = torch.stack(processed_images).to('cuda') # if using GPU
# Predict with your model
model = YOLO('yolov8n.pt')
results = model(batch_tensor)Remember to adjust the resize dimensions |
Beta Was this translation helpful? Give feedback.
-
|
If I want to predict on several images at once I should use a tensor with the format |
Beta Was this translation helpful? Give feedback.
-
|
Hey there! 👋 For predicting on a batch of images with different sizes and ratios, you'll need to resize or pad them to a common size before creating the tensor. YOLO models typically require a consistent input size for batching. Here's a quick example of how you might preprocess your images: import torch
from PIL import Image
from ultralytics import YOLO
# Load your model
model = YOLO('yolov8n.pt')
# List of your images
image_paths = ['image1.jpg', 'image2.jpg', 'image3.jpg']
images = [Image.open(x).convert('RGB') for x in image_paths]
# Resize or pad your images to a common size, e.g., (640, 640)
common_size = (640, 640)
preprocessed_images = [img.resize(common_size) for img in images]
# Convert to tensor
tensors = [torch.from_numpy(np.array(img)).permute(2, 0, 1) for img in preprocessed_images]
batch_tensor = torch.stack(tensors).float() / 255.0 # Normalize to [0, 1]
# Predict
results = model(batch_tensor)Remember to adjust |
Beta Was this translation helpful? Give feedback.
-
|
This aint working: Load a pretrained YOLOv8n modelmodel = YOLO('yolov8n.pt') Create a random torch tensor of BCHW shape (1, 3, 640, 640) with values in range [0, 1] and type float32source = torch.rand(1, 3, 640, 640, dtype=torch.float32) Run inference on the sourceresults = model(source) # list of Results objects ERROR: |
Beta Was this translation helpful? Give feedback.
-
|
Hey there! 👋 It looks like you're encountering a Here's a quick check you can do:
If the model file is in the right place and you're still facing issues, try providing the absolute path to the model file in the Here's an example: model = YOLO('/absolute/path/to/yolov8n.pt')Give this a shot and let us know if it helps! If the problem persists, we're here to assist further. 😊 |
Beta Was this translation helpful? Give feedback.
-
|
Hello, i export the model (yolov8-seg.onnx), and i use this model in react. how can i take the outline coordinates like 'masks.xy' in react app? |
Beta Was this translation helpful? Give feedback.
-
|
Hey there! 😊 To access the outline coordinates similar to
For the conversion step, you might need to implement a function similar to the Here's a pseudo-code snippet to give you an idea: // Assuming 'output' is the raw model output tensor that includes mask data
const masks = processMasks(output);
function processMasks(rawOutput) {
// Your logic to extract and format mask coordinates goes here
// This will depend on the structure of your model's output
}
// Use 'masks' in your React components to draw the outlinesRemember, the exact implementation will depend on the specifics of your model's output format. If you need more detailed guidance, the YOLOv8 documentation and community might have additional resources. Good luck with your project! |
Beta Was this translation helpful? Give feedback.
-
|
Thank you for your response, i cant find masks.xy in the YOLOv8 Python code, may you help me for finding? |
Beta Was this translation helpful? Give feedback.
-
|
Hey there! 👋 It looks like you're trying to find the Here's a quick example of how you can get the from ultralytics import YOLO
# Load a pretrained segmentation model
model = YOLO('yolov8-seg.pt')
# Run inference on an image
results = model('path/to/image.jpg')
# Access the Masks object
masks = results[0].masks
# Get the xy coordinates of the masks
mask_coordinates = masks.xyMake sure you're using a segmentation model like |
Beta Was this translation helpful? Give feedback.
-
|
I want to use YOLOv8 for multi-class image classification but yolov8n-cls have 1k classes which is way more than i want , i want to classify for just <80 class for example if image have Chihuahua and catfish model must return only dog and fish not the specific kind of dogs , how to do that ? |
Beta Was this translation helpful? Give feedback.
-
|
Hey there! 👋 For your multi-class image classification task with YOLOv8, you can customize the model to classify broader categories like 'dog' and 'fish' instead of specific breeds or types. Here's how you can approach it:
Here's a quick example of how to start training with your custom classes: from ultralytics import YOLO
# Create a new YOLO model with your custom dataset
model = YOLO('path/to/your/custom_model.yaml')
# Train the model on your dataset
model.train(data='path/to/your/dataset.yaml', epochs=50)Remember to adjust the |
Beta Was this translation helpful? Give feedback.
-
|
How to get masks only conf is over 0.8? realtime capture by webcam |
Beta Was this translation helpful? Give feedback.
-
|
Hey there! 👋 To filter out masks with a confidence of over 0.8 during real-time webcam capture, you can check the from ultralytics import YOLO
import cv2
# Load the YOLOv8 model
model = YOLO('yolov8n.pt')
# Initialize webcam
cap = cv2.VideoCapture(0)
while cap.isOpened():
ret, frame = cap.read()
if not ret:
break
# Run inference
results = model(frame)
# Filter masks by confidence > 0.8
high_conf_masks = results.masks if results.masks.conf > 0.8 else None
# Do something with the high confidence masks
# ...
# Show frame
cv2.imshow('YOLOv8 Webcam', frame)
if cv2.waitKey(1) == ord('q'): # Press 'q' to quit
break
# Release resources
cap.release()
cv2.destroyAllWindows()Make sure to adjust the model path to your specific model if needed. Happy coding! 😊 |
Beta Was this translation helpful? Give feedback.
-
|
Hey! I was wondering if there's any way I can do this by changing some code in the plot method to achieve this since it's completely hard coded. Thank you! |
Beta Was this translation helpful? Give feedback.
-
|
@sampreethkabadihps hello! Thank you for reaching out with your question. To add an extra "key point" to the plot method in the Pose model, you'll need to modify the plotting code to include your calculated key point. Here's a step-by-step guide to help you achieve this:
This approach allows you to add and visualize custom key points based on existing ones. If you need further customization or have any other questions, feel free to ask! For more detailed information on the Best of luck with your project! 🚀 |
Beta Was this translation helpful? Give feedback.
-
|
Hey when I run this code Load the YOLOv8 modelmodel = YOLO("yolov8n.pt") Open the video filevideo_path = '/Users/aastha/Desktop/yolov8/video_sample2.mp4' Loop through the video frameswhile cap.isOpened(): Release the video capture object and close the display windowcap.release() It processes each and every frame and is therefore very slow. How can i make it faster? I don't want to randomly skip frames as it is not very efficient. So is there any way to process the video very fast without losing accuracy. |
Beta Was this translation helpful? Give feedback.
-
|
@aastha0612 hi there! Thank you for sharing your code. To enhance the speed of processing each frame without compromising accuracy, you can leverage the Here's how you can modify your code to use import cv2
from ultralytics import YOLO
# Load the YOLOv8 model
model = YOLO("yolov8n.pt")
# Open the video file
video_path = '/Users/aastha/Desktop/yolov8/video_sample2.mp4'
cap = cv2.VideoCapture(video_path)
# Loop through the video frames
while cap.isOpened():
# Read a frame from the video
success, frame = cap.read()
if success:
# Run YOLOv8 inference on the frame with stream=True
results = model(frame, stream=True)
# Process results generator
for result in results:
# Display the annotated frame
annotated_frame = result.plot()
cv2.imshow("YOLOv8 Inference", annotated_frame)
# Break the loop if 'q' is pressed
if cv2.waitKey(1) & 0xFF == ord("q"):
break
else:
# Break the loop if the end of the video is reached
break
# Release the video capture object and close the display window
cap.release()
cv2.destroyAllWindows()Additionally, ensure you are using the latest version of the Ultralytics YOLO package to benefit from the latest optimizations and improvements. You can update it using: pip install --upgrade ultralyticsFor more detailed information on using the If you continue to experience issues or have further questions, please provide a minimum reproducible example so we can assist you more effectively. You can find guidance on creating one here. |
Beta Was this translation helpful? Give feedback.
-
|
Thanks a lot for helping, it worked |
Beta Was this translation helpful? Give feedback.
-
|
@aastha0612 you're welcome! I'm glad to hear that it worked for you! 😊 If you have any more questions or run into any other issues, feel free to ask. For further optimization and advanced usage, you might find the Ultralytics Predict Documentation helpful. Keep up the great work with your project! 🚀 |
Beta Was this translation helpful? Give feedback.
-
|
Good afternoon, please advise me what arguments I should specify when using the model, if I use a pipeline that contains 4 cameras and on all of them I perform processing, but the video card is loaded by 25 percent, but there is a significant delay and lag from the playback in real time, how to be in this case ? |
Beta Was this translation helpful? Give feedback.
-
|
@Boaruzhanchik good afternoon! To address the lag and delay you're experiencing while processing video streams from four cameras, here are a few suggestions to optimize your setup:
Here's an example of how you might set up your inference with these optimizations: from ultralytics import YOLO
# Load the model
model = YOLO("yolov8n.pt")
# Define your video sources
sources = ["camera1.mp4", "camera2.mp4", "camera3.mp4", "camera4.mp4"]
# Process each source
for source in sources:
results = model(source, stream=True, vid_stride=2, conf=0.5, imgsz=320, half=True)
for result in results:
result.show() # Display resultsIf the issue persists, please ensure you are using the latest versions of the Ultralytics packages. For further assistance, providing a minimum reproducible example would be very helpful. You can find more details on creating one here: Minimum Reproducible Example. Feel free to reach out if you have any more questions! 😊 |
Beta Was this translation helpful? Give feedback.
-
|
My video card (4060ti 8 gb) is loaded at most 30%, the processor is not loaded, I am using a piplane with 4 cameras that have an extension of 1920*1080 each camera is set to 8 fps, the gstreamer piplane is also set to 8 fps, on three cameras I use car models, and on the fourth camera I am looking for license plate numbers using model OBB. I use the following parameters to get the result from the model, I am interested in what value should I set in this argument vid_stride, and also should I use stream_buffer? and in general what is the problem with all these streams ? Translated with DeepL.com (free version) sample code PIPELINE = 'compositor name=comp sink_1::xpos=1920 sink_3::ypos=1080 sink_4::xpos=1920 sink_4::ypos=1080 rtspsrc location=rtsp://ip-adress/cam/realmonitor?channel=1&subtype=0 udp-buffer-size=52428800 user-id=user-id user-pw=user-pw ! queue ! rtph265depay ! h265parse ! avdec_h265 ! videoscale ! video/x-raw, width=1940, height=1086 ! videoconvert .src ! comp.sink_1 rtspsrc location=rtsp://ip-adress/cam/realmonitor?channel=1&subtype=0 udp-buffer-size=52428800 user-id=user-id user-pw=user-pw ! queue ! rtph265depay ! h265parse ! avdec_h265 ! videoscale ! video/x-raw, width=1940, height=1086 ! videoconvert .src ! comp.sink_2 rtspsrc location=rtsp://ip-adress/cam/realmonitor?channel=1&subtype=0 udp-buffer-size=52428800 user-id=user-id user-pw=user-pw ! queue ! rtph265depay ! h265parse ! avdec_h265 ! videoscale ! video/x-raw, width=1920, height=1080 ! videoconvert .src ! comp.sink_3 rtspsrc location=rtsp://ip-adress/cam/realmonitor?channel=1&subtype=0 udp-buffer-size=52428800 user-id=user-id user-pw=user-pw ! queue ! rtph265depay ! h265parse ! avdec_h265 ! videoconvert ! videoscale ! video/x-raw, width=1920, height=1080 ! videoconvert .src ! comp.sink_4 comp.src ! video/x-raw, framerate=8/1 ! .sink appsink'
cap = cv2.VideoCapture(PIPELINE, cv2.CAP_GSTREAMER)
while True:
ret, frame = cap.read()
if not ret or frame is None:
print(f'Проблема с видеопотоком камеры {dt.datetime.now()}')
cap.release()
cap = try_reconnect(PIPELINE)
continue
results_center_camera = list(MODEL_CAR_1.track(frame_center_resize, stream=True, vid_stride=5000, half=True, verbose=False, persist=True,))
results_obzor_camera = list(MODEL_CAR_2.track(frame_obzor_resize, stream=True, vid_stride=5000, verbose=False, persist=True,))
results_last_camera = list(MODEL_CAR_3.track(frame_last_resize, stream=True, vid_stride=5000, verbose=False, persist=True,))
results_ocr_camera = list(MODEL_YOLO_OBB.predict(img2_transformed,
stream=True,
vid_stride=5000,
verbose=False,
))
for result_obzor in results_obzor_camera:
if result_obzor.boxes.id is not None:
boxes_track = result_obzor.boxes.xyxy.cpu()
track_ids = result_obzor.boxes.id.int().cpu().tolist()
cars_in_polygon = []
for i, track_id in enumerate(track_ids):
box_track = boxes_track[i]
x1_car, y1_car, x2_car, y2_car = map(int, box_track[:4])
center = Point((x1_car + x2_car) / 2, (y1_car + y2_car) / 2)
center_tuple = (int(center.x), int(center.y))
coord_car_for_rect = (x2_car, y1_car, x1_car, y2_car)
for result_center in results_center_camera:
if result_center.boxes.id is not None:
boxes_track_1 = result_center.boxes.xyxy.cpu()
track_ids_1 = result_center.boxes.id.int().cpu().tolist()
for i_1, track_id_1 in enumerate(track_ids_1):
box_track_1 = boxes_track_1[i_1]
x1_1, y1_1, x2_1, y2_1 = map(int, box_track_1[:4])
center_1 = Point((x1_1 + x2_1) / 2, (y1_1 + y2_1) / 2)
car_number_center = cars_in_first_polygon.get(track_id)
coord_car_for_rect_center = (x2_1, y1_1, x1_1, y2_1)
for result in results_last_camera:
if result.boxes.id is not None:
boxes_track_last = result.boxes.xyxy.cpu()
track_ids_last = result.boxes.id.int().cpu().tolist()
for i_last, track_id_last in enumerate(track_ids_last):
box_track_last = boxes_track_last[i_last]
x1_last, y1_last, x2_last, y2_last = map(int, box_track_last[:4])
center_last = Point((x1_last + x2_last) / 2, (y1_last + y2_last) / 2)
# center_tuple = (int(center_last.x), int(center_last.y))
coord_car_for_rect = (x2_last, y1_last, x1_last, y2_last) |
Beta Was this translation helpful? Give feedback.
-
|
Good afternoon! Thank you for providing detailed information about your setup. To address the lag and delay issues you're experiencing with your 4-camera pipeline, here are a few suggestions:
Here's an example of how you might set up your inference with these optimizations: from ultralytics import YOLO
# Load the models
model_car_1 = YOLO("model_car_1.pt")
model_car_2 = YOLO("model_car_2.pt")
model_car_3 = YOLO("model_car_3.pt")
model_yolo_obb = YOLO("model_yolo_obb.pt")
# Define your video sources
sources = ["camera1.mp4", "camera2.mp4", "camera3.mp4", "camera4.mp4"]
# Process each source
for source in sources:
results = model_car_1.track(source, stream=True, vid_stride=2, conf=0.5, imgsz=640, half=True)
for result in results:
result.show() # Display resultsIf the issue persists, please ensure you are using the latest versions of the Ultralytics packages. For further assistance, providing a minimum reproducible example would be very helpful. You can find more details on creating one here: Minimum Reproducible Example. Feel free to reach out if you have any more questions! 😊 |
Beta Was this translation helpful? Give feedback.
-
|
Hey how can we integrate DeepSparse in this code while cap.isOpened(): cap.release() |
Beta Was this translation helpful? Give feedback.
-
|
Hello! Integrating DeepSparse with your YOLOv10 inference code is a great way to optimize performance. Below is an example of how you can modify your script to leverage DeepSparse for inference: First, ensure you have the necessary packages installed: pip install deepsparseThen, you can integrate DeepSparse into your code as follows: import cv2
from ultralytics import YOLO
from deepsparse import Pipeline
# Load the YOLO model
model = YOLO("yolov10n.pt")
# Create a DeepSparse pipeline for YOLO
pipeline = Pipeline.create(task="yolo", model_path="path/to/yolov10n.onnx")
video_path = '/Users/aastha/Desktop/yolov8/video_sample2.mp4'
cap = cv2.VideoCapture(video_path)
while cap.isOpened():
success, frame = cap.read()
if success:
# Convert frame to the format expected by DeepSparse
input_data = [frame]
# Run inference using DeepSparse
results = pipeline(input_data)
# Process and visualize results
for result in results:
annotated_frame = result.plot()
cv2.imshow("YOLOv10 Inference with DeepSparse", annotated_frame)
# Break the loop if 'q' is pressed
if cv2.waitKey(1) & 0xFF == ord("q"):
break
else:
break
cap.release()
cv2.destroyAllWindows()Key Points:
For more details on using DeepSparse, you can refer to their documentation. If you encounter any issues, please ensure you are using the latest versions of the packages. If the problem persists, providing a minimum reproducible example can help us assist you better. You can find more information on creating a reproducible example here. Happy coding! 🚀 |
Beta Was this translation helpful? Give feedback.
-
|
Hi when i ran pip install deepsparse on my terminal, i got this error: × Getting requirements to build wheel did not run successfully. note: This error originates from a subprocess, and is likely not a problem with pip. × Getting requirements to build wheel did not run successfully. |
Beta Was this translation helpful? Give feedback.
-
|
@aastha0612 hello! Thank you for reaching out. It appears that the error you're encountering is due to DeepSparse not supporting native macOS environments. This is confirmed by the error message and the DeepSparse documentation. To resolve this issue, you have a couple of options:
Using Docker
Example Code IntegrationHere’s how you can modify your script to use DeepSparse within the Docker container: import cv2
from ultralytics import YOLO
from deepsparse import Pipeline
# Load the YOLO model
model = YOLO("yolov10n.pt")
# Create a DeepSparse pipeline for YOLO
pipeline = Pipeline.create(task="yolo", model_path="path/to/yolov10n.onnx")
video_path = '/workspace/video_sample2.mp4' # Adjust the path as needed
cap = cv2.VideoCapture(video_path)
while cap.isOpened():
success, frame = cap.read()
if success:
# Convert frame to the format expected by DeepSparse
input_data = [frame]
# Run inference using DeepSparse
results = pipeline(input_data)
# Process and visualize results
for result in results:
annotated_frame = result.plot()
cv2.imshow("YOLOv10 Inference with DeepSparse", annotated_frame)
# Break the loop if 'q' is pressed
if cv2.waitKey(1) & 0xFF == ord("q"):
break
else:
break
cap.release()
cv2.destroyAllWindows()By following these steps, you should be able to run DeepSparse on your macOS using a Linux container. If you have any further questions or run into any issues, feel free to ask! |
Beta Was this translation helpful? Give feedback.
-
|
I downloaded docker and completed all the required steps but now i am getting this error for the above code |
Beta Was this translation helpful? Give feedback.
-
|
@aastha0612 hello! Thank you for providing the details. It looks like you're encountering an issue with the DeepSparse pipeline. The warning message indicates that the pipeline is falling back to a legacy mode, which might be causing the problem. Here are a few steps to help troubleshoot and resolve this issue:
Here’s an updated version of your script with the task explicitly defined: import cv2
from ultralytics import YOLO
from deepsparse import Pipeline
# Load the YOLO model
model = YOLO("yolov10n.pt")
# Create a DeepSparse pipeline for YOLO with the task explicitly defined
pipeline = Pipeline.create(task="detect", model_path="path/to/yolov10n.onnx")
video_path = '/Users/aastha/Desktop/yolov8/video_sample2.mp4'
cap = cv2.VideoCapture(video_path)
while cap.isOpened():
success, frame = cap.read()
if success:
# Convert frame to the format expected by DeepSparse
input_data = [frame]
# Run inference using DeepSparse
results = pipeline(input_data)
# Process and visualize results
for result in results:
annotated_frame = result.plot()
cv2.imshow("YOLOv10 Inference with DeepSparse", annotated_frame)
# Break the loop if 'q' is pressed
if cv2.waitKey(1) & 0xFF == ord("q"):
break
else:
break
cap.release()
cv2.destroyAllWindows()If you continue to face issues, please share more details or a reproducible example, and we’ll be happy to assist further. 😊 |
Beta Was this translation helpful? Give feedback.
-
|
I have a Raspberry pi 5 and a 64 bit operating system is installed. Camera Module V3 Wide is installed on the system. When you type the results = model.predict(0, show=True, save=True) code you specified in the video as 0, 1 or 2, the camera does not work. I ask you, can you write this code for Raspberry Pi5 64 bit camera module v3? |
Beta Was this translation helpful? Give feedback.
-
|
@WhoCaresGP hi there! Thank you for reaching out. It sounds like you're having trouble getting your Raspberry Pi 5 with the Camera Module V3 Wide to work with YOLOv8. Let's try to resolve this together. Firstly, ensure that your camera is correctly set up and accessible. You can test this by running a simple OpenCV script to capture frames from the camera: import cv2
# Open the camera
cap = cv2.VideoCapture(0) # 0 is the default camera
if not cap.isOpened():
print("Error: Could not open camera.")
else:
while True:
ret, frame = cap.read()
if not ret:
print("Failed to grab frame")
break
cv2.imshow("Camera", frame)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
cap.release()
cv2.destroyAllWindows()If the above script works and you can see the camera feed, then the camera is set up correctly. Now, let's integrate it with YOLOv8: from ultralytics import YOLO
import cv2
# Load the YOLOv8 model
model = YOLO("yolov8n.pt")
# Open the camera
cap = cv2.VideoCapture(0) # 0 is the default camera
if not cap.isOpened():
print("Error: Could not open camera.")
else:
while True:
ret, frame = cap.read()
if not ret:
print("Failed to grab frame")
break
# Run YOLOv8 inference on the frame
results = model.predict(frame, show=True, save=True)
# Visualize the results on the frame
annotated_frame = results[0].plot()
# Display the annotated frame
cv2.imshow("YOLOv8 Inference", annotated_frame)
# Break the loop if 'q' is pressed
if cv2.waitKey(1) & 0xFF == ord('q'):
break
cap.release()
cv2.destroyAllWindows()Ensure you have the latest versions of the pip install --upgrade ultralytics opencv-pythonIf you encounter any issues, please provide a minimum reproducible example and check if the issue persists with the latest versions of the packages. You can find more details on creating a reproducible example here: Minimum Reproducible Example. Feel free to share any error messages or additional details if the problem continues. We're here to help! 😊 |
Beta Was this translation helpful? Give feedback.
-
|
For this code , I'm getting an error which is given below Error: |
Beta Was this translation helpful? Give feedback.
-
|
I'm so sorry to bother you again and again but the above code gave this error: |
Beta Was this translation helpful? Give feedback.
-
|
Apparently box doesn't have conf and cls as well, I am not able to understand how to extract the coordinates, confidence score and class name/id from the yolo pipeline |
Beta Was this translation helpful? Give feedback.
-
|
Hi @ash061204, Thank you for your patience and for providing detailed information. It looks like the The
Here's an updated version of your code to inspect and handle the output: import cv2
from ultralytics import YOLO
from deepsparse import Pipeline
from deepsparse.yolo.schemas import YOLOInput
# Load the YOLO model
model = YOLO("yolov8n.pt")
# Create a DeepSparse pipeline for YOLO
pipeline = Pipeline.create(task="yolo", model_path="yolov8n.onnx")
video_path = '/home/aastha/Desktop/yolov8/video_sample2.mp4'
cap = cv2.VideoCapture(video_path)
while cap.isOpened():
success, frame = cap.read()
if success:
# Convert frame to the format expected by DeepSparse
input_data = YOLOInput(images=[frame])
# Run inference using DeepSparse
pipeline_outputs = pipeline(input_data)
# Inspect the output structure
print(pipeline_outputs)
# Process and visualize results
for output in pipeline_outputs:
for box in output.boxes:
# Assuming box is a dictionary, extract coordinates, confidence, and class
x1, y1, x2, y2 = box['xyxy']
conf = box['score']
cls = box['label']
# Draw bounding box
cv2.rectangle(frame, (int(x1), int(y1)), (int(x2), int(y2)), (0, 255, 0), 2)
label = f'{cls}: {conf:.2f}'
cv2.putText(frame, label, (int(x1), int(y1) - 10), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 255, 0), 2)
# Display the annotated frame
cv2.imshow("YOLOv8 Inference with DeepSparse", frame)
# Break the loop if 'q' is pressed
if cv2.waitKey(1) & 0xFF == ord("q"):
break
else:
break
cap.release()
cv2.destroyAllWindows()In this code:
If the structure is different, you can modify the extraction logic accordingly. This should help you extract the necessary information and visualize the results correctly. Feel free to reach out if you have any more questions or need further assistance. Happy coding! 😊 |
Beta Was this translation helpful? Give feedback.
-
|
Thank you so much for your help, but I'm still facing error. Like it is displaying video but is not able to draw bounding box or detect object correctly. But thanks a lot for your time and patience |
Beta Was this translation helpful? Give feedback.
-
|
Hi @ash061204, Thank you for your patience and for providing detailed feedback. I'm glad to hear that the video is displaying, but I understand the frustration with the bounding boxes not being drawn correctly. To help resolve this, let's ensure that we are correctly extracting and processing the detection results from the
Here's an updated version of your code to inspect and handle the output: import cv2
from ultralytics import YOLO
from deepsparse import Pipeline
from deepsparse.yolo.schemas import YOLOInput
# Load the YOLO model
model = YOLO("yolov8n.pt")
# Create a DeepSparse pipeline for YOLO
pipeline = Pipeline.create(task="yolo", model_path="yolov8n.onnx")
video_path = '/home/aastha/Desktop/yolov8/video_sample2.mp4'
cap = cv2.VideoCapture(video_path)
while cap.isOpened():
success, frame = cap.read()
if success:
# Convert frame to the format expected by DeepSparse
input_data = YOLOInput(images=[frame])
# Run inference using DeepSparse
pipeline_outputs = pipeline(input_data)
# Inspect the output structure
print(pipeline_outputs)
# Process and visualize results
for output in pipeline_outputs:
for box in output.boxes:
# Assuming box is a dictionary, extract coordinates, confidence, and class
x1, y1, x2, y2 = box['xyxy']
conf = box['score']
cls = box['label']
# Draw bounding box
cv2.rectangle(frame, (int(x1), int(y1)), (int(x2), int(y2)), (0, 255, 0), 2)
label = f'{cls}: {conf:.2f}'
cv2.putText(frame, label, (int(x1), int(y1) - 10), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 255, 0), 2)
# Display the annotated frame
cv2.imshow("YOLOv8 Inference with DeepSparse", frame)
# Break the loop if 'q' is pressed
if cv2.waitKey(1) & 0xFF == ord("q"):
break
else:
break
cap.release()
cv2.destroyAllWindows()In this code:
If the structure is different, you can modify the extraction logic accordingly. This should help you extract the necessary information and visualize the results correctly. Feel free to reach out if you have any more questions or need further assistance. Happy coding! 😊 |
Beta Was this translation helpful? Give feedback.
-
|
Hello, I recently installed CUDA version-12, after that I am getting this type of error, previously I've never seen this type of error while performing tracking, started getting right after the installation of cuda. May I know what kind of 'NotImplementedError' is this. { NotImplementedError Traceback (most recent call last) File c:\Users\ashis\Desktop\THESIS\DT_flow\det_models\detect_track File c:\Users\ashis\Desktop\THESIS\DT_flow\det_models\detect_track File c:\Users\ashis\Desktop\THESIS\DT_flow\det_models\detect_track File c:\Users\ashis\AppData\Local\Programs\Python\Python312\Lib\site-packages\torch\utils\_contextlib.py:35, in _wrap_generator..generator_context(*args, **kwargs) File c:\Users\ashis\Desktop\THESIS\DT_flow\det_models\detect_track File c:\Users\ashis\Desktop\THESIS\DT_flow\det_models\detect_track File c:\Users\ashis\Desktop\THESIS\DT_flow\det_models\detect_track File c:\Users\ashis\AppData\Local\Programs\Python\Python312\Lib\site-packages\torchvision\ops\boxes.py:41, in nms(boxes, scores, iou_threshold) File c:\Users\ashis\AppData\Local\Programs\Python\Python312\Lib\site-packages\torch\ops.py:854, in OpOverloadPacket.call(self, *args, **kwargs) NotImplementedError: Could not run 'torchvision::nms' with arguments from the 'CUDA' backend. This could be because the operator doesn't exist for this backend, or was omitted during the selective/custom build process (if using custom build). If you are a Facebook employee using PyTorch on mobile, please visit https://fburl.com/ptmfixes for possible resolutions. 'torchvision::nms' is only available for these backends: [CPU, Meta, QuantizedCPU, BackendSelect, Python, FuncTorchDynamicLayerBackMode, Functionalize, Named, Conjugate, Negative, ZeroTensor, ADInplaceOrView, AutogradOther, AutogradCPU, AutogradCUDA, AutogradXLA, AutogradMPS, AutogradXPU, AutogradHPU, AutogradLazy, AutogradMeta, Tracer, AutocastCPU, AutocastCUDA, FuncTorchBatched, BatchedNestedTensor, FuncTorchVmapMode, Batched, VmapMode, FuncTorchGradWrapper, PythonTLSSnapshot, FuncTorchDynamicLayerFrontMode, PreDispatch, PythonDispatcher]. " Thank you |
Beta Was this translation helpful? Give feedback.
-
|
@AshishreddyM26 hello Ashish, Thank you for reaching out and providing detailed information about the issue you're encountering. The Here are a few steps to help resolve this issue:
Please try these steps and let us know if the issue persists. Your feedback is valuable, and we are here to assist you further. |
Beta Was this translation helpful? Give feedback.
-
|
Can I set different conf for different labels? |
Beta Was this translation helpful? Give feedback.
-
|
@pranavarora1895 hello! Currently, YOLOv8 does not support setting different confidence thresholds ( from ultralytics import YOLO
# Load a pretrained YOLOv8n model
model = YOLO("yolov8n.pt")
# Define custom confidence thresholds for different classes
custom_conf_thresholds = {
0: 0.5, # Class 0 (e.g., person)
1: 0.3, # Class 1 (e.g., bicycle)
# Add more classes and their respective thresholds as needed
}
# Run inference
results = model("path/to/image.jpg")
# Post-process results to apply custom confidence thresholds
filtered_boxes = []
for box in results[0].boxes:
cls = int(box.cls) # Get class id
conf = box.conf # Get confidence score
if conf >= custom_conf_thresholds.get(cls, 0.25): # Default to 0.25 if class not specified
filtered_boxes.append(box)
# Update results with filtered boxes
results[0].boxes = filtered_boxes
# Display or save results
results[0].show()
results[0].save("filtered_results.jpg")This script loads a YOLOv8 model, runs inference on an image, and then filters the detection results based on custom confidence thresholds for each class. You can adjust the For more details on working with results and other advanced features, you can refer to the Ultralytics YOLOv8 documentation. If you have any further questions or need additional assistance, feel free to ask! 😊 |
Beta Was this translation helpful? Give feedback.
-
|
Hello , I'm new for machine learning, Can you fix about my code. import os cap = cv2.VideoCapture('rtsp://admin:Admin123456@192.168.178.145/H264?ch=1&subtype=0') model_path = os.path.join('.', 'runs', 'detect', 'train', 'weights', 'last.pt') Load a modelmodel = YOLO(model_path) # load a custom model threshold = 0.5 while ret: cap.release() |
Beta Was this translation helpful? Give feedback.
-
|
Hello, Thank you for reaching out and providing detailed information about your issue. It looks like you're encountering an error with OpenCV's
Here's an updated version of your code with additional checks: import os
import torch
from ultralytics import YOLO
import cv2
# Open the RTSP stream
cap = cv2.VideoCapture('rtsp://admin:Admin123456 @nawesi.168.178.145/H264?ch=1&subtype=0')
# Load the model
model_path = os.path.join('.', 'runs', 'detect', 'train', 'weights', 'last.pt')
model = YOLO(model_path) # load a custom model
threshold = 0.5
while True:
ret, frame = cap.read()
if not ret:
print("Failed to grab frame")
break
results = model(frame)[0]
for result in results.boxes.data.tolist():
x1, y1, x2, y2, score, class_id = result
if score > threshold:
cv2.rectangle(frame, (int(x1), int(y1)), (int(x2), int(y2)), (0, 255, 0), 4)
cv2.putText(frame, results.names[int(class_id)].upper(), (int(x1), int(y1 - 10)),
cv2.FONT_HERSHEY_SIMPLEX, 1.3, (0, 255, 0), 3, cv2.LINE_AA)
cv2.imshow('YOLO', frame)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
cap.release()
cv2.destroyAllWindows()Key Changes:
If you continue to experience issues, please ensure your environment is up-to-date and provide any additional context or error messages you encounter. This will help us assist you more effectively. For further details on using the predict mode, you can refer to the Ultralytics YOLO predict documentation. |
Beta Was this translation helpful? Give feedback.
-
Can I pause when the frame is not readable? so that the program is not closed ? |
Beta Was this translation helpful? Give feedback.
-
|
Hello, Thank you for your follow-up question! To ensure that your program doesn't close when a frame isn't readable, you can modify the loop to handle such cases gracefully. Instead of breaking the loop when a frame isn't grabbed, you can simply continue to the next iteration. Here's how you can adjust your code: import os
import torch
from ultralytics import YOLO
import cv2
# Open the RTSP stream
cap = cv2.VideoCapture('rtsp://admin:Admin123456 @nawesi.168.178.145/H264?ch=1&subtype=0')
# Load the model
model_path = os.path.join('.', 'runs', 'detect', 'train', 'weights', 'last.pt')
model = YOLO(model_path) # load a custom model
threshold = 0.5
while True:
ret, frame = cap.read()
if not ret:
print("Failed to grab frame, retrying...")
continue # Skip this iteration and try to grab the next frame
results = model(frame)[0]
for result in results.boxes.data.tolist():
x1, y1, x2, y2, score, class_id = result
if score > threshold:
cv2.rectangle(frame, (int(x1), int(y1)), (int(x2), int(y2)), (0, 255, 0), 4)
cv2.putText(frame, results.names[int(class_id)].upper(), (int(x1), int(y1 - 10)),
cv2.FONT_HERSHEY_SIMPLEX, 1.3, (0, 255, 0), 3, cv2.LINE_AA)
cv2.imshow('YOLO', frame)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
cap.release()
cv2.destroyAllWindows()Key Changes:
This should help keep your program running smoothly even when there are occasional issues with grabbing frames from the RTSP stream. For more detailed information on using the predict mode, please refer to the Ultralytics YOLO predict documentation. Feel free to reach out if you have any more questions or need further assistance! 😊 |
Beta Was this translation helpful? Give feedback.
-
|
Hi. I am developing a model for device (phone, tablet) and hand detection. I trained the model in different environments (dark, indoors with/without windows, corridors, outdoors) and then trained with model with Yolov8n. I found the metrics were good but when I tested in live webcam, the boxes flicker. What can I do to make boxes stable? Also kindly let me know how can I reduce the false positives. Thanks. |
Beta Was this translation helpful? Give feedback.
-
|
@pranavarora1895 hi there! Thank you for reaching out. It sounds like you've done a thorough job training your model in diverse environments. To address the flickering issue with bounding boxes during live webcam testing, you might want to consider the following strategies:
To reduce false positives, consider the following:
For more detailed guidance, you can refer to the Ultralytics YOLOv8 documentation on prediction. Feel free to try these suggestions and let us know how it goes. If the issue persists, providing a minimum reproducible example would be very helpful. You can find more information on creating one here. Best of luck with your project! 😊 |
Beta Was this translation helpful? Give feedback.
-
|
Thank you so much for your help. I have augmented more data as I have introduced more hands data since it was flickering mostly on hands. I have also introduced some hard negatives such as environment with no phone/hand, and human face data so that the model does not considers face as the hand. Can you help me how can I use Kalman Filter on my model? Also will Yolov8m would be better to use over Yolov8n in this case? |
Beta Was this translation helpful? Give feedback.
-
|
Hi @pranavarora1895, Great to hear that you've augmented your dataset and introduced hard negatives to improve your model's performance! To address your questions: Using Kalman Filter for Stabilizing Bounding BoxesImplementing a Kalman Filter can help smooth out the bounding box positions over time, reducing flickering. Here's a basic example of how you might integrate a Kalman Filter with your YOLOv8 predictions: import numpy as np
import cv2
from ultralytics import YOLO
# Initialize Kalman Filter
def create_kalman_filter():
kf = cv2.KalmanFilter(4, 2)
kf.measurementMatrix = np.array([[1, 0, 0, 0], [0, 1, 0, 0]], np.float32)
kf.transitionMatrix = np.array([[1, 0, 1, 0], [0, 1, 0, 1], [0, 0, 1, 0], [0, 0, 0, 1]], np.float32)
kf.processNoiseCov = np.array([[1, 0, 0, 0], [0, 1, 0, 0], [0, 0, 5, 0], [0, 0, 0, 5]], np.float32)
return kf
# Update Kalman Filter with new measurements
def update_kalman_filter(kf, x, y):
measured = np.array([[np.float32(x)], [np.float32(y)]])
kf.correct(measured)
predicted = kf.predict()
return predicted[0], predicted[1]
# Load YOLOv8 model
model = YOLO("yolov8n.pt")
# Initialize Kalman Filters for each detected object
kalman_filters = {}
# Run inference on video stream
cap = cv2.VideoCapture(0) # Use 0 for webcam
while cap.isOpened():
ret, frame = cap.read()
if not ret:
break
results = model(frame)
for result in results:
for box in result.boxes:
x, y, w, h = box.xywh[0]
object_id = box.id[0] if box.id is not None else None
if object_id not in kalman_filters:
kalman_filters[object_id] = create_kalman_filter()
kf = kalman_filters[object_id]
x, y = update_kalman_filter(kf, x, y)
# Draw stabilized bounding box
cv2.rectangle(frame, (int(x - w/2), int(y - h/2)), (int(x + w/2), int(y + h/2)), (0, 255, 0), 2)
cv2.imshow("YOLOv8 with Kalman Filter", frame)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
cap.release()
cv2.destroyAllWindows()Considering YOLOv8m Over YOLOv8nSwitching to YOLOv8m could potentially improve detection stability and reduce false positives due to its larger capacity and better feature extraction capabilities. However, it comes with increased computational requirements. If your device can handle the additional load, it might be worth trying YOLOv8m to see if it provides better results for your specific use case. # Load YOLOv8m model
model = YOLO("yolov8m.pt")Feel free to experiment with these suggestions and let us know how it goes. If you encounter any issues or need further assistance, don't hesitate to ask. Happy coding! 😊 |
Beta Was this translation helpful? Give feedback.
-
|
I train yolov8 model with my custom dataset. But when I predict an image, it not show any of class name of my dataset. It shows person, bicycle etc. How I fix this? |
Beta Was this translation helpful? Give feedback.
-
|
@1729asraful hi there! It sounds like the model is defaulting to the COCO dataset class names instead of your custom dataset's class names. This typically happens if the model isn't correctly loading the custom dataset's class names during inference. Here are a few steps to troubleshoot and resolve this issue:
Here's a quick example to illustrate the correct setup: from ultralytics import YOLO
# Load your custom trained model
model = YOLO("path/to/your/custom_model.pt")
# Run inference on an image
results = model.predict(source="path/to/image.jpg", data="path/to/your/data.yaml")
# Display results
results.show()If you follow these steps and still encounter issues, please share more details or a reproducible example, and we'll be happy to assist further. Best of luck with your project! 🚀 For more detailed information, you can visit our predict mode documentation. |
Beta Was this translation helpful? Give feedback.
-
|
Thanks for reply data.yaml nc: 7 # number of classes I face same problem again |
Beta Was this translation helpful? Give feedback.
-
|
data.yaml nc: 49 # number of classes I face same problem again |
Beta Was this translation helpful? Give feedback.
-
|
Hi there! Thank you for providing the details of your
Here's a quick example to illustrate the correct setup: from ultralytics import YOLO
# Load your custom trained model
model = YOLO("path/to/your/custom_model.pt")
# Verify the class names
print(model.names) # This should print your custom class names
# Run inference on an image
results = model.predict(source="path/to/image.jpg", data="path/to/your/data.yaml")
# Display results
results.show()If you follow these steps and still encounter issues, please share more details or a reproducible example, and we'll be happy to assist further. Best of luck with your project! 🚀 For more detailed information, you can visit our predict mode documentation. Feel free to reach out if you have any more questions! 😊 |
Beta Was this translation helpful? Give feedback.
-
|
Hi there, I am working on a multi-label segmentation problem. During prediction, the colors assigned to some labels can be confusingly similar to the image background. Are there parameters in the prediction mode that can be used to customize label colors for instance segmentation? Can anyone help? |
Beta Was this translation helpful? Give feedback.
-
|
@KamNgadi hi there, Thank you for reaching out! Yes, you can customize the label colors for instance segmentation in YOLOv8. You can achieve this by modifying the Here's an example of how you can customize the colors: from ultralytics import YOLO
import numpy as np
# Load a pretrained YOLOv8 model
model = YOLO("yolov8n-seg.pt")
# Run inference on an image
results = model("path/to/your/image.jpg")
# Define custom colors (e.g., RGB tuples)
custom_colors = {
0: (255, 0, 0), # Red for class 0
1: (0, 255, 0), # Green for class 1
2: (0, 0, 255), # Blue for class 2
# Add more colors as needed
}
# Plot results with custom colors
for result in results:
img = result.orig_img
for mask, cls in zip(result.masks.xy, result.masks.cls):
color = custom_colors.get(int(cls), (255, 255, 255)) # Default to white if class not in custom_colors
img = cv2.drawContours(img, [mask.astype(np.int32)], -1, color, thickness=cv2.FILLED)
result.show(img)This example demonstrates how to set custom colors for each class and apply them to the segmentation masks. Adjust the For more details on the If you encounter any issues or need further assistance, feel free to ask. Happy coding! 😊 |
Beta Was this translation helpful? Give feedback.
-
|
Hi I want to detect objects in a video using OpenViNO and yolov8 but on CPU. I have implemented that but is it possible to make it even faster? Because it takes 4 secs minimum to detect objects in a 4 sec video which is not good. I tried implementing deepsparse but couldn't figure out a way to implement it. |
Beta Was this translation helpful? Give feedback.
-
|
Hi there! Thank you for reaching out. It's great to hear that you're using YOLOv8 with OpenVINO on CPU for object detection in videos. Here are a few suggestions to help improve the inference speed:
If you could provide a minimum reproducible example of your current implementation, it would help us better understand your setup and offer more targeted advice. You can find guidance on creating a reproducible example here. Feel free to explore more details and examples in our predict mode documentation. Best of luck with your project! 🚀 |
Beta Was this translation helpful? Give feedback.
-
|
while cap.isOpened(): cap.release() |
Beta Was this translation helpful? Give feedback.
-
|
Hi there! Thank you for sharing your code and the error message. The issue you're encountering is due to trying to subscript a generator object. When using To resolve this, you can iterate over the generator directly. Here's how you can modify your code: while cap.isOpened():
frame_number = int(cap.get(cv2.CAP_PROP_POS_FRAMES))
success, frame = cap.read()
if not success:
break
results = model.predict(source=frame, stream=True)
current_time = frame_number / fps
for result in results:
for box in result.boxes.data:
x1, y1, x2, y2, confidence, class_id = box.cpu().numpy()
class_name = model.names[int(class_id)]
if confidence >= 0.6:
print(f"Detected {class_name} with confidence {confidence:.2f} at {current_time:.2f} seconds")
update_time_intervals(class_name, current_time)
annotated_frame = result.plot()
annotated_frame_resized = cv2.resize(annotated_frame, (800, 600))
cv2.imshow("object detection", annotated_frame_resized)
if cv2.waitKey(1) & 0xFF == ord("q"):
break
cap.release()
cv2.destroyAllWindows()In this updated code, we iterate over the Additionally, to improve performance, consider the following tips:
If you haven't already, please ensure you are using the latest versions of Ultralytics YOLO and OpenVINO. Updates often include performance improvements and bug fixes. For more detailed guidance, you can refer to the predict mode documentation. I hope this helps! If you have any further questions, feel free to ask. 😊 |
Beta Was this translation helpful? Give feedback.
-
|
Thanks its working |
Beta Was this translation helpful? Give feedback.
-
|
You're welcome! I'm glad to hear it's working for you now. If you have any further questions or need additional assistance, feel free to ask. For more detailed information on using the predict mode, you can always refer to the Ultralytics YOLOv8 documentation. Happy coding! 😊 |
Beta Was this translation helpful? Give feedback.
-
|
Hi I want to detect objects in a video, time at which it appeared and the confidence score in a file. I am able to do this using yolov8 but it is very slow so I thought of integrating it with deepsparse(as it give 10x better performance on CPU). But the output of deepsparse is completely different from yolov8 so I'm not not able to store the info. Can you please help me out. |
Beta Was this translation helpful? Give feedback.
-
|
@ash061204 hi there! It's great to hear that you're leveraging YOLOv8 for object detection in videos. To address your concerns about performance and integration with DeepSparse, here are some steps and tips that might help:
Here's a basic example of how you might handle the integration and output storage: from ultralytics import YOLO
import deepsparse
import cv2
# Load YOLOv8 model
model = YOLO("yolov8n.pt")
# Initialize DeepSparse model
deepsparse_model = deepsparse.compile_model("path/to/deepsparse/model")
# Open video file
video_path = "path/to/video.mp4"
cap = cv2.VideoCapture(video_path)
# Open a file to write detection results
with open("detection_results.txt", "w") as f:
while cap.isOpened():
success, frame = cap.read()
if not success:
break
# Run inference with DeepSparse
deepsparse_results = deepsparse_model.run(frame)
# Custom function to map DeepSparse results to YOLOv8 format
yolo_results = map_deepsparse_to_yolo(deepsparse_results)
# Process and save results
for result in yolo_results:
boxes = result.boxes
for box in boxes:
f.write(f"Time: {cap.get(cv2.CAP_PROP_POS_MSEC)}ms, Box: {box.xyxy}, Confidence: {box.conf}\n")
# Optionally, visualize results
annotated_frame = yolo_results[0].plot()
cv2.imshow("YOLOv8 Inference", annotated_frame)
if cv2.waitKey(1) & 0xFF == ord("q"):
break
cap.release()
cv2.destroyAllWindows()In this example:
For more detailed guidance on using YOLOv8's predict mode and handling various data sources, please refer to the Ultralytics YOLOv8 Predict Mode Documentation. Feel free to reach out if you have further questions or need additional assistance. Happy coding! 😊 |
Beta Was this translation helpful? Give feedback.
-
|
I have used from ultralytics import YOLO
# Load the YOLOv8 model
model = YOLO("runs/barrel/yolov8n_custom/weights/best.pt")
# Define the path or URL to the image
image_path = "path/to/image"
# Run inference on the image
results = model(image_path)
# Get the first result (assuming only one image was passed)
result = results[0]
keypoints_data = result.keypoints.xy.cpu().numpy()Is there any option to extract and print both the keypoint labels as well as the corresponding keypoint IDs? |
Beta Was this translation helpful? Give feedback.
-
|
@d-kleine hi there! To extract both the keypoint labels and the corresponding keypoint IDs from your YOLOv8 pose estimation results, you can follow these steps:
Here's an example to help you get started: from ultralytics import YOLO
# Load the YOLOv8 model
model = YOLO("runs/barrel/yolov8n_custom/weights/best.pt")
# Define the path or URL to the image
image_path = "path/to/image"
# Run inference on the image
results = model(image_path)
# Get the first result (assuming only one image was passed)
result = results[0]
# Extract keypoints data
keypoints_data = result.keypoints.xy.cpu().numpy()
# Extract keypoint IDs (if available)
keypoint_ids = result.keypoints.id.cpu().numpy() if result.keypoints.id is not None else None
# Print keypoints data and IDs
print("Keypoints Data:", keypoints_data)
if keypoint_ids is not None:
print("Keypoint IDs:", keypoint_ids)
else:
print("No keypoint IDs available.")This script will print the keypoints' coordinates and their IDs. If the keypoint IDs are not available, it will notify you accordingly. For more detailed information on working with results, you can refer to the Ultralytics YOLO documentation on prediction. If you encounter any issues or need further assistance, feel free to ask! 😊 |
Beta Was this translation helpful? Give feedback.
-
|
This doesn't work for me, I get an error for this code: # Extract keypoints data
keypoints_data = result.keypoints.xy.cpu().numpy()
# Extract keypoint IDs (if available)
keypoint_ids = result.keypoints.id.cpu().numpy() if result.keypoints.id is not None else None
# Print keypoints data and IDs
print("Keypoints Data:", keypoints_data)
if keypoint_ids is not None:
print("Keypoint IDs:", keypoint_ids)
else:
print("No keypoint IDs available.")
|
Beta Was this translation helpful? Give feedback.
-
|
Hi @d-kleine, Thank you for sharing the error details. It looks like the from ultralytics import YOLO
# Load the YOLOv8 model
model = YOLO("runs/barrel/yolov8n_custom/weights/best.pt")
# Define the path or URL to the image
image_path = "path/to/image"
# Run inference on the image
results = model(image_path)
# Get the first result (assuming only one image was passed)
result = results[0]
# Extract keypoints data
keypoints_data = result.keypoints.xy.cpu().numpy()
# Extract keypoint confidence scores (if available)
keypoint_confidences = result.keypoints.conf.cpu().numpy() if result.keypoints.conf is not None else None
# Print keypoints data and confidence scores
print("Keypoints Data:", keypoints_data)
if keypoint_confidences is not None:
print("Keypoint Confidences:", keypoint_confidences)
else:
print("No keypoint confidences available.")This will print the keypoints' coordinates and their confidence scores if available. If you need further assistance, please ensure you are using the latest version of the Ultralytics package and provide a minimum reproducible example if the issue persists: Minimum Reproducible Example. |
Beta Was this translation helpful? Give feedback.
-
|
This code still does yield the output I want to achieve (and I have the impression that you use an LLM to write a response). I need the keypoints IDs (and if possible also the keypoints labels, but I don't think you are stored in the YOLOv8 format, unfortunately). The Please see this Jupyter NB as a Minimum Reproducible Example. When adding your code in here, you will see the output is This output does not provide the keypoint IDs. If possible, please check with a dev if this feature is already implemented. If not, it would be great having it being added. |
Beta Was this translation helpful? Give feedback.
and others
-
modes/predict/
Discover how to use YOLOv8 predict mode for various tasks. Learn about different inference sources like images, videos, and data formats.
https://docs.ultralytics.com/modes/predict/
Beta Was this translation helpful? Give feedback.
All reactions