Nerogar
I'm dyslexic with spelling issues can captioning have the ability to add from txt
Describe your use-case.
request: Anyway with my issues dyslexic and spelling, i have a txt document with key words to my styles etc I always use, I copy and past the text into every txt made from the auto captioning tool from that txt file, this is time consuming, I would love it if it would be possible to have a selection add from txt file where we can select add new line and it automatically do this incorporating the text from the selected text file.
not sure it be handy for may but id definitely use it. thank you
What would you like to see as a solution?
add new line to txt from txt import file in captioning
Have you considered alternatives? List them here.
No response
I'm not quite sure what exactly you mean. But if you want to add a prefix to all generated captions, you should use the "caption prefix" field instead of the "initial caption" field. Initial caption is sent to the model to continue generating. Prefix is just added to the start of all captions.
The requirements are not clear and OT isnt a captioning software. Go use a tool dedicated for this like, its the best one I have so far found:
https://github.com/Particle1904/DatasetHelpers
I'm not quite sure what exactly you mean. But if you want to add a prefix to all generated captions, you should use the "caption prefix" field instead of the "initial caption" field. Initial caption is sent to the model to continue generating. Prefix is just added to the start of all captions.
i do its broken for wd14 only not blips
The requirements are not clear and OT isnt a captioning software. Go use a tool dedicated for this like, its the best one I have so far found:
https://github.com/Particle1904/DatasetHelpers
one trainer has captioning built in, wd14 has a small issue that im pointing out here, it has blip and blip2 captioning but need to run blip first if you go strait to blip2 it will error as does pull blip required part just blip 2 update parts but they work.
and im requests is for a add to that captioning from a text file as it has the feature to add cations to the already made captions, using its built in captioning
The requirements are not clear and OT isnt a captioning software. Go use a tool dedicated for this like, its the best one I have so far found: https://github.com/Particle1904/DatasetHelpers
one trainer has captioning built in, wd14 has a small issue that im pointing out here, it has blip and blip2 captioning but need to run blip first if you go strait to blip2 it will error as does pull blip required part just blip 2 update parts but they work.
and im requests is for a add to that captioning from a text file as it has the feature to add cations to the already made captions, using its built in captioning
Read the name - One Trainer.
The requirements are not clear and OT isnt a captioning software. Go use a tool dedicated for this like, its the best one I have so far found: https://github.com/Particle1904/DatasetHelpers
one trainer has captioning built in, wd14 has a small issue that im pointing out here, it has blip and blip2 captioning but need to run blip first if you go strait to blip2 it will error as does pull blip required part just blip 2 update parts but they work. and im requests is for a add to that captioning from a text file as it has the feature to add cations to the already made captions, using its built in captioning
Read the name - One Trainer.
you have me lost with your comment, yes one trainer as in has everything in it, as in one solution
i mean im dyslexic with reading writing issues, but what im reading is use another tool not the built in features
I'm not quite sure what exactly you mean. But if you want to add a prefix to all generated captions, you should use the "caption prefix" field instead of the "initial caption" field. Initial caption is sent to the model to continue generating. Prefix is just added to the start of all captions.
[ initial caption ] doesn't add, So say i add, XARTX at initial caption the generated text does not have it using wd14
it is working for blip and blip 2
No, initial caption does not add it to the caption. Some of the models, like BLIP and BLIP2 continue generating text from an initial text. WD14 does not do this, it just generates a list of tags. If you want to add something to a generated WD14 caption, use the prefix instead of the initial caption.
This is not a bug, it's just how the models work
No, initial caption does not add it to the caption. Some of the models, like BLIP and BLIP2 continue generating text from an initial text. WD14 does not do this, it just generates a list of tags. If you want to add something to a generated WD14 caption, use the prefix instead of the initial caption.
This is not a bug, it's just how the models work
ok thank for that, i thought prefex was the seperator the ,
joys for dyslexic lol
thank you that is a massive help.
Note still would love add from text, due to description words i use for my scanned in art work, that i have to do manually all the time.
i use wd14 then blip to add a second line with that description works well for my needs, I'm loving onetrainer way easyer for me than kohya
I told you what to use. Stop being a vexatious, spamming user and go try it. That captioning software literally solves your problem as it supports mass adds, resorting and prioritising of words and many many more. More captioning models too.
Additionally consider that you could also try out TagGUI.
I told you what to use. Stop being a vexatious, spamming user and go try it. That captioning software literally solves your problem as it supports mass adds, resorting and prioritising of words and many many more. More captioning models too.
Additionally consider that you could also try out TagGUI.
i looked at it you seem to miss I have dyslexic and autism the program is not intuitive, its not spamming its a request. what your doing is bulling. the feature set of onetrainer is fantastic and my request doesn't have to be do, all I'm doing is asking. its my first request also, how is that spamming. I'd honestly appreciate if you could please leave me alone.
Not sure to fully catch your request : " I would love it if it would be possible to have a selection add from txt file where we can select add new line". Is that always the same text with your tags ? In that case, there is a possible workaround: Instead of adding manually this text as a new line in each caption, you could just duplicate your concept and select the option "from single text file" as prompt source. Then play around the repeats (<1) for your concepts if you don't want to double the image number.
I mean one concept with generated prompts and option "From text file per sample" and another with the option "from single text file" where you select your unique text file. Then optionally use the balancing option for each concepts if you don't want to double the image numbers.
I mean one concept with generated prompts and option "From text file per sample" and another with the option "from single text file" where you select your unique text file. Then optionally use the balancing option for each concepts if you don't want to double the image numbers.
i mean i have a text fill with, oil painting, brush marks, wild colors, thick palette knife style, vivid etc etc that i copy and past into at current 109 text files to each of my paintings,
id love a option just to select that txt file so when i use wd14 to also adds my text saving me open copy past repeatedly.
also helpful with a character where i am doing this method also.
Well, honnestly I still don't really catch the request but I still think some workaround can be found with the existing options. You can use several options for a concept for caption, from single text file or from a single text file? You also can duplicate concepts, choose other caption option, play with balancing.