Re: [tesseract-ocr] How to get the net_spec

[Shree Devi Kumar]

https://github.com/tesseract-ocr/tessdoc/blob/main/Data-Files-in-tessdata_best.md

https://github.com/tesseract-ocr/tessdoc/blob/main/Data-Files-in-tessdata_fast.md

Version string : 4.00.00alpha : [Network specification] for tessdata_best

tessdata_best models - incomplete list, only till Kannada.

The flags are TrainingFlags from lstmrecognizer.h. 0x40 is compress unicharset and 1 is integer mode. The one from best has flags 40 = compress + not integer mode.

afr
Version string:4.00.00alpha:afr:synth20170629:[1,36,0,1Ct3,3,16Mp3,3Lfys64Lfx96Lrx96Lfx512O1c1]
LSTM training info:Network str:[1,36,0,1Ct3,3,16Mp3,3Lfys64Lfx96Lrx96Lfx512O1c1], 
flags=40, iteration=286700, sample_iteration=286724, null_char=95, 
learning_rate=0.001, momentum=0.5, adam_beta=0.999

amh
Version string:4.00.00alpha:amh
LSTM training info:Network str:[1,36,0,1Ct3,3,16Mp3,3Lfys48Lfx96Lrx96Lfx192O1c1], 
flags=40, iteration=6112200, sample_iteration=6112270, null_char=284, 
learning_rate=0.001, momentum=0.5, adam_beta=0.999

On Fri, Sep 15, 2023, 9:50 PM Des Bw <desal...@gmail.com> wrote:

For the last couple of days, I have been trying to train the amh data to include some missing characters. 

I have seen that Shree was able to add the Norwegian Æ by removing the top layer and training on it (https://groups.google.com/g/tesseract-ocr/c/l33zsTEPj70/m/wPzPv6HiEQAJ). 

I was trying to do the same. But, the traineddata in Amharic doesn't contain the net_spec information with the version line. 

Version:4.00.00alpha:amh:synth20170629

17:lstm:size=3356155, offset=192

18:lstm-punc-dawg:size=3154, offset=3356347

19:lstm-word-dawg:size=5007810, offset=3359501

20:lstm-number-dawg:size=810, offset=8367311

21:lstm-unicharset:size=18906, offset=8368121

22:lstm-recoder:size=2578, offset=8387027

23:version:size=30, offset=8389605

Can sb  (@Shree, please) help me on how to get the net_spec, or how to proceed to add a layer to introduce the missing characters?

--
You received this message because you are subscribed to the Google Groups "tesseract-ocr" group.
To unsubscribe from this group and stop receiving emails from it, send an email to tesseract-oc...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/tesseract-ocr/5bedab52-2f9b-44ab-a97a-e2033d0e92den%40googlegroups.com.

[Shree Devi Kumar]

The language name headings seem to be missing from the tessdoc page for tessdata_fast

Please revert to an older version of page from history 

[Shree Devi Kumar]

combine_tessdata(1) is the main program to combine/extract/overwrite/list/compact tessdata components in [lang].traineddata files.

Options to list network details

-d .traineddata FILE...: Lists directory of components from the .traineddata file.

-l .traineddata FILE...: List the network information.

[Des Bw]

Thank you so much dear Shree. You are a life saver. 

[Tom Morris]

On Saturday, September 16, 2023 at 4:11:48 PM UTC-4 shree wrote:

The language name headings seem to be missing from the tessdoc page for tessdata_fast

Please revert to an older version of page from history 

 

I've regenerated the tessdata_fast page using the latest `combine_tessdata` and included the language names.

https://github.com/tesseract-ocr/tessdoc/pull/120

I also regenerated the tessdata_best page so that it no longer is truncated at Kannada. Both pages now separate the models for languages and scripts to reflect the new directory structure.

Tom

[Tom Morris]

On Saturday, September 16, 2023 at 4:25:51 PM UTC-4 shree wrote:

combine_tessdata(1) is the main program to combine/extract/overwrite/list/compact tessdata components in [lang].traineddata files.

Options to list network details

-d .traineddata FILE...: Lists directory of components from the .traineddata file.

-l .traineddata FILE...: List the network information.

It was undocumented (until now), but you can also combine the above as -dl or -ld to get both sets of information at once.

Tom 

[Des Bw]

Thank you for adding those improvements dear Tom.