Caffe convert_imageset with one class/ label - neural-network

I want to create an lmdb dataset from images which part of them contain the feature I want caffe to learn, and part of them don't.
My question is - in the text input file transferred to convert_imageset - how should I label those images that don't contain the feature?
I know the format is
PATH_TO_IMAGE LABEL
PATH_TO_IMAGE LABEL
PATH_TO_IMAGE LABEL
But which label should I assign to images without the feature?
For example, img1.jpg contain the feature, img2.jpg and img3.jpg don't.
So should the text file look like -
img1.jpg 0
img2.jpg 1?
img3.jpg 1?
Thanks!

Got an answer from Caffe-users Google Group - yes, creating a dummy feature is the right way for this.
So it is:
img1.jpg 0
img2.jpg 1
img3.jpg 1

Related

How to train/fine tune with images containing multiple languages?

When extracting wordstrboxes I supply the languages the image contains, for example "eng+deu", however lstmf is created per language.
tesseract.exe image.png image --psm 4 -l eng+deu wordstrbox
So what is the correct path forward here? Use the images both in training for English and for German? Or something else?
Thanks

Ag-Grid Not Supporting Special Symbol

I tried using Ag-Grid, and faced a challenge to plot the data for Special Symbol, It actually doesn't support the special symbol and doesn't plot it!!
Refer To Example1.png, here the No. Column Doesn't contain data,
but when I remove the dot operator from my script.js file from headerName,
and rowData, it starts plotting it. (Refer To Example2.png)
Also attached a zip file (Google Drive Link) as a prototype of this problem for your reference!
https://drive.google.com/open?id=1UqyrRtAhg8-HBwqVqUT_CGMhwyr_cOUt
Kushagra
You probably just need to set the suppressFieldDotNotation grid option to true:
suppressFieldDotNotation
If true, then dots (eg address.firstline) in field names are not treated as deep references. Allows you to use dots in your field name
if you prefer.
(You didn't show any code and I'm not looking in a .zip.)
Alternatively, you can have dots in your headerName, they just can't be in your field name.

Extract word between two words in Tableau

My data in a single column looks like this
Environment: PROD_A
JobName: MY JOB NAME 1
MemName: Some_Script_names
My aim is to create a calculated field where it fetches JobName.
In above sample. The output should be MY_JOB_NAME_1
Here's my code in Tableau -
MID([Description],FIND([Description],"JobName: ")+9,
FIND([Description],"MemName: ")-FIND([Description],"JobName: ")-9)
The above logic is same as excel function below. It basically finds first word and last word and then picks the one in between.
=MID(A1,SEARCH("JobName: ",A1)+9,SEARCH("MemName",A1)-SEARCH("JobName: ",A1)-9)
Issue -
Tableau is showing blank data. Not sure where my logic is wrong. Any help is appreciated.
You may be able to use REGEX funtions (depending on your data source). Try this: REGEXP_EXTRACT([Data],'JobName: (.*)')
I see that you commented that the text is in one cell and not one column as you wrote in question.
This finds the JobName and uses that position plus the lenght of it as the startposition, then finds MemName as the endposition (and calculates the number of characters between them)
=MID(A1,FIND("JobName: ",A1)+LEN("JobName: "),FIND("MemName",A1,FIND("JobName: ",A1)+LEN("JobName: ")+1)-(FIND("JobName: ",A1)+Len("JobName: ")))
The following tableau calc works, this assumes that there is no space after MY_JOB_NAME_1 and "MemName:"
MID([Text],
FIND([Text],'JobName: ')+10,
FIND(
MID([Text],
FIND([Text],'JobName: ')+10)," ")-9)

MATLAB: datasets stuck in each box

I'm unsure how to phrase exactly my problem. I have run a script and it is putting each file of data into one box, making four files, a 1x4 array. I can click on each and t will expand, however this is not the format that is acceptable. Is it possible to extract this to turn it into one single file? I attached a picture.
Your second file has 23 columns instead of 24, so you cannot just add them together. If you add a column to the second file you can convert them using:
b=cell2mat(a')
adding a column can be done like this:
a{2}(:,24)=nan;

Jasper reports: Overlapped fields not shown in excel output

Take the following example:
I have one very wide column (lets say 150pt), positioned on x=0
I have 2 columns of 25pt, positioned on x=100 and x=125. Thus, these are overlapping the first.
Depending on certain conditions (parameters to the report), I do or do not print the 2 overlapping columns. I do this by using the "print when expression ...".
This works like a charm when I use the PDF as output, but when I generate the report in excel, I do not get the big field, it is just missing. As long as I do not print the 2 overlapping fields, everything remains OK.
Any ideas on how to solve this one?
Thanks
Sounds like the answer is no ... hopefully there is a better answer
http://community.jaspersoft.com/questions/503288/missing-columns-excel
... the columns are overlapping. And the so-called "grid
exporters" like the HTML, XLS and CSV exporters do not support
overlappging elements. The elements that are behind do not print.
I hope this helps. Teodor