-
Notifications
You must be signed in to change notification settings - Fork 130
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Only one model in the repository? #237
Comments
For Arabic there is https:/OpenITI/OCR_GS_Data, there is the legacy https:/mittagessen/kraken-models and there are various older @pontelneptique shared a Latin mlmodel here: mittagessen/kraken-models#5 |
Thanks. |
* Janusz S. Bień :: 2021-01-25 12:43 Mon:
Moreover I don't know how to use a model which is not listed by 'kraken list' :-(
You just use the -m option like for repository models and put the path
of the model file instead. The model repository is empty as its use is
not documented and requires some technical skill to get the identifiers
for zenodo and such. It is going to be added to escriptorium soon-ish
though which should improve the situation.
In the meanwhile there are people keeping models in various git
repositories...
|
Thanks again. |
You can see the definition of the CLI in |
Go to https:/mittagessen/kraken-models/blob/ec4407159e19d0df6e366038f262b4f5a8bbb12a/mlmodel/latin-teubner/latin-teubner.mlmodel and click "Download" |
* Janusz S. Bień :: 2021-01-25 12:57 Mon:
Please note that `kraken --help` doesn't show `-m` option. What should I consult to avoid asking elementary questions?
It is an option on the ocr subcommand. `kraken ocr --help` should print
it (and a million others).
There is some documentation here [0] on the option usage for the
different subcommands, although that's for the last stable release and
not the development branch.
[0] http://kraken.re/advanced.html
|
Thank you very much! |
Hi, I'm having real trouble understanding the documentation and parts of this thread. |
* Fergus :: 2021-08-06 20:32 Fri:
Hi, I'm having real trouble understanding the documentation and parts of this thread.
Can you please tell me what the command would be to install a custom model from a Github repo?
If I can just download the file and put it in the Kraken directory, then which directory and where?
Yes, you can just download the model file and place it whereever you
want. Then you can use it by pointing kraken to the file with the `-i`
or `-m` option depending on the command you're actually running. The
auto-finding feature that scours the default directory is completely
optional.
|
I have tried using the -m option, and am still having no luck. I have the following models downloaded on my PC and I am also trying to get one from GitHub. Please advise on exactly which commands I should use. I appreciate your help.
|
@FergusJPWalsh If you already have the models downloaded, you don't need to call |
But when I call
|
Could I ask that the whole process of installing a new model be explained to me, step by step and with the appropriate commands? |
Let's say yout want to process wget https:/pharos-alexandria/ocr-greek_cursive/raw/master/kraken-models/model_grc_savile.mlmodel
kraken -i foo.png binarize segment -bl ocr -m model_grc_savile.mlmodel This will download the model and run binarization, segment and ocr with kraken. |
Thank you so much. That's been a great help. |
I just installed kraken and
kraken list
gives only10.5281/zenodo.2577813 (pytorch) - A generalized model for English printed text
.Where are other models? I'm especially interested in Medieval Latin.
The text was updated successfully, but these errors were encountered: