Whenever i try to run yolonas model in my jupyter notebook i always get error no module found for super gradient even after i have pip install super-gradients too how to solve this in jupyter notebook? Do we need to install additional library?
hi I want to run this detection on raspberry pi it is possible? After getting best files how to train in raspberry pi or jetson nano can you give solution @@LearnOpenCV
Can anyone please how to save the model after training on custom dataset, so that I don't have to train again and again for inference in different type of videos
YOLOv8 and YOLO NAS are two different detection architectures. Although parts of the architectures do partially share some common properties, model conversion actually doesn't make sense here. However the repositories for both of the frameworks do have provision to export models across various frameworks (PyTorch, TF lite, ONNX, etc) for inference.
@@LearnOpenCV thank you so much for the help and answer, right now i am trying to convert YOLOv6 model with ONNX type to TensorRT (engine) but still it sucks because I can't install CUDA in my windows laptop, do you have any advice to convert that without CUDA?
Would you recommend object tracking for tracking a syringe in a botox training simulator? Or do you think the tech isn’t quite there for such small visual changes?
Hi, how to filter out a specific class in YOLO-nas eg. if I want to detect only 'Person', what should be modified here out = model.predict("/content/1.jpg", conf=0.25)
Hey, you should check the output variable. Check the code at this link for more information: github.com/Deci-AI/super-gradients/blob/master/YOLONAS.md#extract-bounding-boxes
great guide , can anyone help me visualize the results like the p pr f1 score and the loss graphs other yolo models used to automatically generate the graphs after training but not yolo nas
I have a YOLO NAS model for animal detection. I ran the model for 25 epochs and have got the best.pth weights. I need to add more epochs, to train it more from where i left off. I have read somewhere that YOLO V5 have such an option. Does YOLO NAS have the same option? If so how can i implement it in colab? PS. It took me 10 - 15 hours to train for 25 epochs. So I am tight on time. I am not sure whether I am doing something wrong, but I am training using A100 GPU in Colab and its taking this much time. Please advice. I have 17.8 GB of data which has around 38790 images, so i guess it makes sense to take that much time? I tried looking through the YOLO NAS documentation and google searched it, but couldn't get any concrete ideas.
When I try to install super-gradients==3.1.1, I keep getting the following error: Failed to build pycocotools ERROR: Could not build wheels for pycocotools, which is required to install pyproject.toml-based projects Anybody know how to solve this? Thanks.
how to import dataset from roboflow to train YOLO NAS, it would be great if you release a video on comparative analysis of YOLO models as you always do
Hi, we are planning YOLO models comparison soon! You can check roboflow's blog post on finetuning the NAS models. blog.roboflow.com/yolo-nas-how-to-train-on-custom-dataset/
Hi, use DDP model of training for multi GPUs. Check this for more info: docs.deci.ai/super-gradients/documentation/source/device.html#3-dp-data-parallel
Whenever i try to run yolonas model in my jupyter notebook i always get error no module found for super gradient even after i have pip install super-gradients too how to solve this in jupyter notebook? Do we need to install additional library?
Make sure you restart the runtime after installing super-gradients!
Great! How to resume to training process if it pauses.
Hey, to resume training, input the last model to the function.
How to get the live feed of the camera after training is completed?
Please refer to this section in this article: learnopencv.com/reading-and-writing-videos-using-opencv/#read-from-web-cam
hi I want to run this detection on raspberry pi it is possible? After getting best files how to train in raspberry pi or jetson nano can you give solution
@@LearnOpenCV
Is this any better than using Ultralytics Yolov8? That is what I use, very easy and less coding.
@evanshlom1 ok. Thanks!
Can anyone please how to save the model after training on custom dataset, so that I don't have to train again and again for inference in different type of videos
All the model checkpoints are saved in their respective "experiment-name" directories.
Hi, can i convert YOLOv8 model to YOLO NAS model
YOLOv8 and YOLO NAS are two different detection architectures. Although parts of the architectures do partially share some common properties, model conversion actually doesn't make sense here.
However the repositories for both of the frameworks do have provision to export models across various frameworks (PyTorch, TF lite, ONNX, etc) for inference.
@@LearnOpenCV thank you so much for the help and answer, right now i am trying to convert YOLOv6 model with ONNX type to TensorRT (engine) but still it sucks because I can't install CUDA in my windows laptop, do you have any advice to convert that without CUDA?
you given the best files of the yolo nas. now i want to run in raspberry pi or jetson nano can you give the procedure or any suggestion for these.
We are working on this. Will share once it is ready.
Would you recommend object tracking for tracking a syringe in a botox training simulator? Or do you think the tech isn’t quite there for such small visual changes?
Yes its possible to do that.
@@LearnOpenCV Thank you!
Can you provide the code to generate the graphs for training the model or can you explain how did you generated the graphs.
The graphs are automatically generated by super-gradients. Check their github repo for more info.
Hi, how to cite YOLO NAS??
Hi, check their GitHub page for citing: opencv-ai-courses-radiologyai.streamlit.app/
Hi, how to filter out a specific class in YOLO-nas eg. if I want to detect only 'Person',
what should be modified here out = model.predict("/content/1.jpg", conf=0.25)
Hey, you should check the output variable. Check the code at this link for more information: github.com/Deci-AI/super-gradients/blob/master/YOLONAS.md#extract-bounding-boxes
great guide , can anyone help me visualize the results like the p pr f1 score and the loss graphs other yolo models used to automatically generate the graphs after training but not yolo nas
Some libraries have inbuilt plotting utilities. If you have the data, you can use matplotlib or seaborn to plot.
I have a YOLO NAS model for animal detection. I ran the model for 25 epochs and have got the best.pth weights. I need to add more epochs, to train it more from where i left off. I have read somewhere that YOLO V5 have such an option. Does YOLO NAS have the same option? If so how can i implement it in colab?
PS. It took me 10 - 15 hours to train for 25 epochs. So I am tight on time. I am not sure whether I am doing something wrong, but I am training using A100 GPU in Colab and its taking this much time. Please advice. I have 17.8 GB of data which has around 38790 images, so i guess it makes sense to take that much time?
I tried looking through the YOLO NAS documentation and google searched it, but couldn't get any concrete ideas.
Would be possible to run YOLO NAS on edge devices like Jetson Nano?
It should not be an issue if using the INT8 version of the YOLO NAS small model. However, running the large model may not give a very high FPS.
@@sovitrath4735 Do you have any recommendation for a tutorial around this?
I can not find where the graphs are generated, even in their main github repo, i could not find any instructions where i can find the training graphs
Hi there. The training we carried out generated tensorboard logs and the graph screenshot is that of the tensforboard dashboard.
Whenever I try to train the YOLO-NAS model on colab , the instance automatically restarts , can you suggest any solution?
Hi, it might be RAM issue. Keep an eye on the resources used tab.
from where to download this dataset
Please check the article on where to download it from: learnopencv.com/train-yolo-nas-on-custom-dataset/
When I try to install super-gradients==3.1.1, I keep getting the following error:
Failed to build pycocotools
ERROR: Could not build wheels for pycocotools, which is required to install pyproject.toml-based projects
Anybody know how to solve this? Thanks.
I change the command to "pip install super-gradients". Now it is fine.
how to import dataset from roboflow to train YOLO NAS, it would be great if you release a video on comparative analysis of YOLO models as you always do
Hi, we are planning YOLO models comparison soon! You can check roboflow's blog post on finetuning the NAS models. blog.roboflow.com/yolo-nas-how-to-train-on-custom-dataset/
@@LearnOpenCV Thank you for the reply will see to it, and I'm waiting for the comparative analysis.
How to use multiple GPUs for training?
Hi, use DDP model of training for multi GPUs. Check this for more info: docs.deci.ai/super-gradients/documentation/source/device.html#3-dp-data-parallel