i'm doing my M.Tech project using orange 3 . all my faculty was amazed by features of orange , thanks to orange 3 for making my project simple and effective.
"Category" is not showing up as an option in my image viewer. Do you have any suggestions as to why? And my import images widget is showing "No data output".
How can we test the model using images the model has not seen? I have tried uploading new photos (in categories as with the training data), used image embedding to it, connected it to the predictions widget, and finally connected the logistic regression widget (which was used to train the model) to the predictions widget . But when I open the predictions widget there are no predictions in them. Can you help please?
Hi, I have a question. Why these flower images already have categories after imported to the workshop? (In the video, it shows "82 images / 9 categories"). Is there a way to assign categories to images? So I can create my own image training set? Thank you.
Categories are based on folder names. As these flowers are in different folders, they are recognized as separate categories. You can do the same for your data. If you don't want categories, just put all images in the same folder.
There show error when put Test & Score Widget to set the target variable. Although I set the target variable through Column widget but the error didn't remove.
In the main folder, you can create sub-folders which are taken as "Categories" and then the error disappears... but It can be argued that this is cheating...
Thank you for the reply. I'll try it out. On a separate note, I think orange is one of the best softwares in this category at the moment. I've recommended it to most of my friends working in this discipline. Please keep the videos coming. It's very much appreciated! God bless.
Orange is mostly dedicated to teaching machine learning without programming and our videos reflect that. You can, however, find some code snippets on our blog.
Thank you for the clear explanation, but unfortunately some image embedding models such as Inception V3 as well as others we cannot use them even with a high speed of internet connection, still I do not know what is the issue?
Hi everyone, I have created a classification model for covid & healthy patient through image analytics. After building the model how can i check whether the new image being entered is correctly classified and how to enter it
Great videos, well done! Is it possible to apply the predictor widget to image embedding and logistic regression (so to predict the class probabilities) ? If yes could give me an example of flows? Thanks!
Sure, just pass the embedded images to Predictions via LogReg (Import Images - Image Embedding - LogReg - Predictions) and then connect the data you are trying to predict to Predictions (New Data - Image Embedding - Predictions).
Orange team, help!! Can you guys explain more regarding topic modeling? I'm confused between LDA, LSI and Hierarchical Dirichlet Process. Thanks in advance xx
How can we test the model using images the model has not seen? I have tried uploading new photos (in categories as with the training data), used image embedding to it, connected it to the predictions widget, and finally connected the logistic regression widget (which was used to train the model) to the predictions widget . But when I open the predictions widget there are no predictions in them. Can you help please?
![[Live สด] การออกรางวัลสลากกินแบ่งรัฐบาล งวดวันที่ 16 ธันวาคม 2567](http://i.ytimg.com/vi/dNTiBBthKgQ/mqdefault.jpg)
i'm doing my M.Tech project using orange 3 .
all my faculty was amazed by features of orange , thanks to orange 3 for making my project simple and effective.
It's amazing & easy tools for data analysis. And It will be more helpful to link dataset used this lecture. Thanks Orange teams.
Awesome! Where can i download the folder "flowers"?
Anyone know how to get the weights and bias for the nodes in a neural network developed in Orange? Thanks
Please provide the dataset you are using in this video. It's not available on the URL specified.
"Category" is not showing up as an option in my image viewer. Do you have any suggestions as to why? And my import images widget is showing "No data output".
How can we test the model using images the model has not seen? I have tried uploading new photos (in categories as with the training data), used image embedding to it, connected it to the predictions widget, and finally connected the logistic regression widget (which was used to train the model) to the predictions widget . But when I open the predictions widget there are no predictions in them. Can you help please?
Could you reach out in our Discord channel and attach a screenshot with the workflow?
Hi, I have a question.
Why these flower images already have categories after imported to the workshop? (In the video, it shows "82 images / 9 categories").
Is there a way to assign categories to images? So I can create my own image training set?
Thank you.
Categories are based on folder names. As these flowers are in different folders, they are recognized as separate categories. You can do the same for your data. If you don't want categories, just put all images in the same folder.
@@OrangeDataMining thank you so much
Is there anyway to get the execution time for the classification algorithm ? how fast does it work?
There show error when put Test & Score Widget to set the target variable.
Although I set the target variable through Column widget but the error didn't remove.
In the main folder, you can create sub-folders which are taken as "Categories" and then the error disappears... but It can be argued that this is cheating...
Hi. Great videos. Is there a link to download the image folder used in this video?
No, as we do not own the images. You can find your own images, create a folder and experiment with that.
Thank you for the reply. I'll try it out. On a separate note, I think orange is one of the best softwares in this category at the moment. I've recommended it to most of my friends working in this discipline. Please keep the videos coming. It's very much appreciated! God bless.
how can I apply gradient descent on images?
Going better and better : a good practical known subject to practice. Can we hope some subject about some python code in a further Video ?
Orange is mostly dedicated to teaching machine learning without programming and our videos reflect that. You can, however, find some code snippets on our blog.
I found the snippets here: blog.biolab.si/tag/python/ pretty small, but ok as a starting point. thanks for the answer
guys can i use this software for remote sensed image classification?
Thank you for the clear explanation, but unfortunately some image embedding models such as Inception V3 as well as others we cannot use them even with a high speed of internet connection, still I do not know what is the issue?
have you solved the issue cause I have the same one
Really awesome 👍
Hi everyone,
I have created a classification model for covid & healthy patient through image analytics. After building the model how can i check whether the new image being entered is correctly classified and how to enter it
Where can I find the flower dataset?
It comes with Orange when you download it. The file is called iris.tab.
Open File widget and you will be able to see it.
Hi, i am susing Orange when image Embading connect to Test & Score widget disply to error signal✖
Me too :(
mee too
same here
Me too
Is it possible to convert the model made using Orange to code. Can it generate code?
No, there's no option to produce code in Orange. People on GitHub have tried different approaches, but as far as I know no one has managed it.
Is it possible to set target variable manually while working on images ?
Yes. You can use Select Columns widget to set the target variable.
Great videos, well done! Is it possible to apply the predictor widget to image embedding and logistic regression (so to predict the class probabilities) ? If yes could give me an example of flows? Thanks!
Sure, just pass the embedded images to Predictions via LogReg (Import Images - Image Embedding - LogReg - Predictions) and then connect the data you are trying to predict to Predictions (New Data - Image Embedding - Predictions).
@@OrangeDataMining GREAT, you should make another tutorial about this important step
That's pretty cool, but how do we apply it to data and not to images? Can you help me with that?
Orange team, help!! Can you guys explain more regarding topic modeling? I'm confused between LDA, LSI and Hierarchical Dirichlet Process. Thanks in advance xx
I am getting this error. Embedding error: SSLError
Hi, is it GPU accelerated?
i don't think so. when doing the image embedding, my CPU usage spike really high while my GPU remains relatively idle
Very Helpful
why you do not talking about machine learning algorithm
muitobon
How can we test the model using images the model has not seen? I have tried uploading new photos (in categories as with the training data), used image embedding to it, connected it to the predictions widget, and finally connected the logistic regression widget (which was used to train the model) to the predictions widget . But when I open the predictions widget there are no predictions in them. Can you help please?