kraken

kraken copied to clipboard

kraken 5.2.9 segmentation training no lines

7

I try to train a segmentation model with `page-xml` data, at start the segmenter shows me the regions and line types but when using the model no lines is detected...

johnlockejrr

Support for Python 3.12

8

When I was trying to add kraken package to my project on Python 3.12 I got error: ``` poetry add kraken Using version ^5.2.5 for kraken Updating dependencies Resolving dependencies......

dantetemplar

Is there a published trained recognition model I can use for Japanese documents?

3

The default model is for english and I have checked the list of available models, but I don't think I found anything for recognition of Japanese documents? Do I have...

BingWangEd

Problem with assigning a line to a region if it's too close to the region border

2

I'm running kraken segmentation inference with the following command, including two different models for baselines- and regions- recognition: ``` kraken -r -d cuda:0 -x -I "$KRAKEN_IN" -o "$KRAKEN_OUT" segment -bl...

fattynoparents

I'm not sure if is `kraken` related or `eScriptorium` related. I want to segment a page in Syriac serto script with vowels so I draw lines as close to the...

johnlockejrr

This PR improves validation metrics for segmentation training by limiting the pixel metrics to regions, vectorizing baselines, and matching them to ground truth baselines.

mittagessen

[Not an Issue] How to keep the trained model as close as possible to groundtruth

3

I try to train a segmentation model for a modern printed Judaeo-Arabic dataset. The problem I face is that in the trained model I mainly loose vowel signs below the...

johnlockejrr

RuntimeError: Caught RuntimeError in pin memory thread for device 0

19

Specs: ``` kraken, version 5.2.10.dev2 Python 3.11.10 11th Gen Intel(R) Core(TM) i7-1165G7 @ 2.80GHz NVIDIA GeForce RTX 3060 12Gb ``` Training small amount of data, about 35 pages. Worked on...

johnlockejrr

Is there an easy way to transfer a model's predictions back into fine-tuning html?

1

Hello, I was wondering if there was a convenient way to take the text obtained from applying a model to an image and insert the lines of that text into...

megamattc

Unable to use PDF processing functions due to missing library

1

Hello, I am running kraken on ubuntu 24.04 with python 3.10 in a conda environment. I am unable to apply kraken models to a pdf document to produce OCR-ed text,...

megamattc