Look at tesstrain.log for details of your training run.
https://github.com/tesseract-ocr/tesseract/wiki/TrainingTesseract-4.00#lstmtraining-command-line
target_error_rate double 0.01 Stop training if the mean percent error rate
gets below this value.
On Fri, Jul 19, 2019 at 11:39 AM Pooja Kamr
PSM 7 was a partial solution for my specific case, it improved the
situation but did not solve it. Also I could not use it in some other cases.
The proper solution is very likely doing more training with more data, some
data augmentation might probably help if data is scarce.
Also doing less train
Lorenzo,
We both have got the same case. It seems a solution to this problem would
save a lot of people.
Shree,
I pulled the current head of master branch but it doesn't seem to contain
the merges you pointed that have been merged 3 to 4 days ago. How can I
get them?
ElMagoElGato
2019年7月1
Is there any way to pass bounding boxes to use to the LSTM? We have an
algorithm that cleanly gets bounding boxes of MRZ characters. However the
results using psm 10 are worse than passing the whole line in. Yet when we
pass the whole line in we get these phantom characters.
Should PSM 10 mode wor
>Is there any way to pass bounding boxes to use to the LSTM?
See
https://github.com/tesseract-ocr/tesseract/wiki/APIExample#getcomponentimages-example
On Fri, Jul 19, 2019 at 2:50 PM Claudiu wrote:
> Is there any way to pass bounding boxes to use to the LSTM? We have an
> algorithm that cleanl
>I pulled the current head of master branch but it doesn't seem to contain
the merges you pointed that have been merged 3 to 4 days ago. How can I
get them?
I usually do `git pull origin master` to get all latest changes from the
master branch.
On Fri, Jul 19, 2019 at 2:35 PM ElGato ElMago
wrot
Lorenzo,
I haven't been checking psm too much. Will turn to those options after I
see how it goes with bounding boxes.
Shree,
I see the merges in the git log and also see that new
option lstm_choice_amount works now. I guess my executable is latest
though I still see the phantom character.
Thanks for that link. Should passing a full image and calling SetRectangle
have different behavior than passing just the cropped image and not using
SetRectangle?
On Fri, Jul 19, 2019 at 11:29 AM Shree Devi Kumar
wrote:
> >Is there any way to pass bounding boxes to use to the LSTM?
>
> See
> htt
As per log file, finished error rate is 1.439.
On Friday, July 19, 2019 at 1:24:35 PM UTC+5:30, shree wrote:
>
> Look at tesstrain.log for details of your training run.
>
>
> https://github.com/tesseract-ocr/tesseract/wiki/TrainingTesseract-4.00#lstmtraining-command-line
>
>
> target_error_rat
Tesseract does not require GPU. But if my system has GPU then will it help
to improve performance.
On Friday, June 28, 2019 at 7:02:30 PM UTC+5:30, Timothy Snyder wrote:
>
> I think it means that Tesseract doesn't support nor require hardware
> acceleration via the GPU.
>
> Looks like there is
As per your screenshot 15000 iterations have been done.
On Fri, Jul 19, 2019 at 3:52 PM Pooja Kamra wrote:
> As per log file, finished error rate is 1.439.
>
>
>
> On Friday, July 19, 2019 at 1:24:35 PM UTC+5:30, shree wrote:
>>
>> Look at tesstrain.log for details of your training run.
>>
>>
>>
I went and tried to understand the source code as well as I could and
although I did not find all the answers I did find some. (for tesseract
4.0.0-beta.3)
At iteration 14615/695400/698614, Mean rms=0.158%, delta=0.295%, char train=
1.882%, word train=2.285%, skip ratio=0.4%, wrote checkpoint.
I went and tried to understand the source code as well as I could and
although I did not find all the answers I did find some. (for tesseract
4.0.0-beta.3)
At iteration 14615/695400/698614, Mean rms=0.158%, delta=0.295%, char train=
1.882%, word train=2.285%, skip ratio=0.4%, wrote checkpoint.
I was confused about the triple iteration number too...
https://groups.google.com/d/msg/tesseract-ocr/hni4owhU3vs/ankF3gSrAwAJ
Le vendredi 19 juillet 2019 12:22:16 UTC+2, Pooja Kamra a écrit :
>
> As per log file, finished error rate is 1.439.
>
>
>
> On Friday, July 19, 2019 at 1:24:35 PM UTC+5
If I understand correctly, you can use Tesseract OCR on GPU for speeding up
the process but not Tesseract training.
Le vendredi 28 juin 2019 07:54:07 UTC+2, Pooja Kamra a écrit :
>
> On Tesseract site, it is mentioned that no GPU is needed (No support).
> What does this statement means?
> If i ha
Very well written. You may want to update the wiki pages with the info too.
On Fri, Jul 19, 2019 at 7:45 PM Arno Loo wrote:
> I went and tried to understand the source code as well as I could and
> although I did not find all the answers I did find some. (for tesseract
> 4.0.0-beta.3)
> At itera
16 matches
Mail list logo
