Using Machine Learning with Detectron2
ฝัง
- เผยแพร่เมื่อ 19 ก.ย. 2024
- Have you ever thought to yourself "Hey I want to do some machine learning, what do I need to know?" and then after some extensive reading you think "Huh, maybe I should just go get a PhD in ML".
Asking yourself all these questions can be overwhelming, and in ML it is sometimes easy to fall down a rabbit hole. But in this video, Cami appeals to those who really just want to get something up and running with ML as quickly as possible. She walks you through how to discover projects that make ML easier, and writes a small application utilizing Detectron2!
------ Social Accounts ------
PyTorch Ecosystem: pytorch.org/ec...
Detectron2 GitHub: github.com/fac...
Cami's Twitter: / cwillycs
P.S. We're hiring! www.surveymonk...
Facebook Open Source Twitter: / fbopensource
Facebook Open Source Facebook Page: / fbopensource
Facebook Open Source website: opensource.face...
------ Facebook Developer Circles ------
Facebook Developer Circles Facebook Page: / developercircles
Facebook Developer Circles GitHub: github.com/fbd...
![[FULL EP.16] เซียนพาลุย "เผือก-ฟรอยด์" กรี๊ดลั่น จนปูหนี | เฮ็ดอย่างเซียนหรั่ง | One Playground](http://i.ytimg.com/vi/FsqUvwS3b-Y/mqdefault.jpg)
This is an awesome tutorial. I would really love to see you make a front-end to it. Thank you.
That was good! Please do more of these Cami!
Cool! Recently I have been searching for "how to make a quick pytorch model api?" so that I can deploy it in any server and use it anywhere!
I had NO idea Colaboratory exited..
Thanks for showing that :)
this is awesome! well done. would love to see the front end part.
Cami
Is it possible to convert Detectron2 as a Rom and run it on different system like Windows or Android?
Will make it into rom allow it to be more compatible?
Thank you very much ! That was awesome, and so are you !!!
Thanks this is a great tutorial!
I was also trying to deal with masks that come from the detectron2 object detector on instance of people; in my case I wanted to 'erase' people from an image of a scene, using the mask. I ended up using the cv2 logical AND operations to convert from the detectron masks to make a binary cv2 image from them and clip out that region from myimage and insert the same masked region taken from a previously taken 'unoccupied' image of the scene with no people in it.
# get mask
pmask = outputs["instances"].pred_masks[pidx].cpu()
nmask = pmask.numpy().astype("uint8")*255
nmask_inv = cv2.bitwise_not(nmask)
cutout = cv2.bitwise_and(myimage, myimage, mask=nmask_inv)
shadow = cv2.bitwise_and(unoccupied, unoccupied, mask=nmask)
removed = cv2.add(cutout, shadow)
cv2_imshow(removed)
how can we get approx size of the object? or is there any way to do it? can it be done while labelling the images if yes then how?
Thanks for the video :) Is there any example of integration of different datasets to detectron-2?
I am working detectron2 object detection model, which produced a good output result. I can also see the output predicted result with labels with the help of detectron2.utils.visualizer Visualizer functions. But I want to print the predicted labels and save them in txt or csv file in the direction from top left side. can you help me with this?
Thank you
Does somebody know how to get the mask from stuff classes? pred_mask only works for instances but not for panoptic_seg
hey, great video !
That was refreshing. 👍
How to use multi-gpu in detectron2 for custom data training?
Funny that this video was added to the page you are showing at 3:26
Hi Guys, I was wondering if any one can help find an easy and simple way of manipulating the prediction object to draw abounding box area the area that has changes in an image.
CUDA error: device-side assert triggered
CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
For debugging consider passing CUDA_LAUNCH_BLOCKING=1.
Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.
i got error like this, what should i do?thank you
Can I install Detectron 2 on windows 10?
I want to do real time object detection
How can we build our own backbone model to embed in this tensorflow object detection api?
And which files we have to update for this purpose?
Let me allow to explain my question, let suppose I introduce my backbone model "A".
In which files I have to give link of "A"? and
Where I have to place "A.py" in object detection api?
Hi. I need to detect circles in a custom model. I have annonated the training images using labelme. It gives 2 no pairs in the json. Can you pls help in how to modify the code for circle detection. Its giving an error as not polygon with current code. Thanks
Is it trained to detect and mask some limited set of specific objects (e.g. only humans, horses and umbrellas for example) or any?
the method closest_color gives me an error... does someone know why?
How to load model.pth in detecreon2 to detection another image? If i was trained the model long time ago?
great video, how to calculate the area of mask region, any guidance please
I tried a couple of the models from the model zoo in the repo, and some detect more objects than the others. Is it possible to do ensembling with this model?
Great video! Anyone know what chrome extension showing time and progress at 07:14
0:41-0:51. Wow I feel called out.
awesome where is the part 2?
tried it could follow up untill 22 minutes than shit hit the fans and couldn't make it work. tried copying the code from the end of the video but doesn't seem to work. would have been better if there was a link to the code.
thanks for the great tutorial
I have a problem
TypeError: 'NoneType' object is not subscriptable
I comes from visualizer = Visualizer(img[:, :, ::-1], metadata=data_metadata, scale=0.5)
Whys the api doc of detectron2 so bad?
Love your content ❤️❤️❤️
Dose detectron2 runs on rtx3090 ?
any one help me for student identification through detectron2 .create dataset
Export model?
wooowo ya da Best
awesome indeed... great thanks for all this channel opens more possibilities ahead... kudos!
roses are red violets are u hehe :)
Hello,
Thank you for the amazingly well explained tutorial. I wrote a code and am using detectron2 to identify objects. It is working great with my in built webcam but whenever I try and use my webcam, it displays only the first frame and then says ..."ASSERT: "false" in file qasciikey.cpp, line 501"
My Hardware config is as follows
Ubuntu 18.04
Nvidia GeForce GTX 1650Ti Card
Torch Version '1.8.1+cu102'
Cuda 10.2
Python 3.6.9
Any help will be much appreciated.
Thanks in advance.
Also my code is unable to read any key inputs when it is running, when attempting to quit the code with
if ord=='q':
cv2.destroyAllWindows()
it either hangs or dumps the code with error
Segmentation fault, core dumped
Thanks
Standard video production error: you have cut away the pauses after each sentence, making the video extremely hard to listen to. Will you please uncut the pauses!
ajjajajajaja
It kills me when u say those topics so easily like nothing
Gradient descend
Back propagation
Neural network
It took me months and i still don't understand them very well 😪
Lol u can say it too it doesnt mean you understand shit.
the most sexiest tutorial on programming so far
What the hell is wrong with you
I don't know where come the sexiest thing is, such a poor guy.
It's not sexiest tutorial.. It's a interesting tutorial
🤣, every child is unique 👍
wow, man when did the word "sexiest" become taboo 😅
i never expected a beverly hills chick to explain me about deep learning
Its LA downtown