How to Use

[aghitshek]

Hello, How can i use this....?
i have pdf of equations and want to convert to LaTeX?
can you give me step by step instruction?
i dont understand in readme

Thank You

[lukas-blecher]

The goal of this project is to convert image snippets of equations to latex code.
So you would have to take a screen shot (eg using the gui) of each equation in your pdf.
Regarding the installation, you will have to follow the readme.
But it sould be simple

1. install python
2. install pytorch
3. install pix2tex[gui]

[lukas-blecher]

1. Did you install pytorch first?
2. There is no pix2tex[cli]. please run pip install pix2tex for client only

[aghitshek]

Yes i did by this command....
pip install torch

[lukas-blecher]

Maybe have a look at this:
pytorch/pytorch#574 (comment)

[aghitshek]

Now when i ran
pix2tex
i get

E:>pix2tex
Traceback (most recent call last):
File "C:\Users\Anonymous\AppData\Local\Programs\Python\Python310\lib\runpy.py", line 196, in _run_module_as_main
return run_code(code, main_globals, None,
File "C:\Users\Anonymous\AppData\Local\Programs\Python\Python310\lib\runpy.py", line 86, in run_code
exec(code, run_globals)
File "C:\Users\Anonymous\AppData\Local\Programs\Python\Python310\Scripts\pix2tex.exe_main.py", line 7, in
File "C:\Users\Anonymous\AppData\Local\Programs\Python\Python310\lib\site-packages\pix2tex_main.py", line 28, in main
main(arguments)
File "C:\Users\Anonymous\AppData\Local\Programs\Python\Python310\lib\site-packages\pix2tex\cli.py", line 224, in main
model = LatexOCR(arguments)
File "C:\Users\Anonymous\AppData\Local\Programs\Python\Python310\lib\contextlib.py", line 79, in inner
return func(*args, **kwds)
File "C:\Users\Anonymous\AppData\Local\Programs\Python\Python310\lib\site-packages\pix2tex\cli.py", line 90, in init
self.image_resizer.load_state_dict(torch.load(os.path.join(os.path.dirname(self.args.checkpoint), 'image_resizer.pth'), map_location=self.args.device))
File "C:\Users\Anonymous\AppData\Local\Programs\Python\Python310\lib\site-packages\torch\serialization.py", line 705, in load
with _open_zipfile_reader(opened_file) as opened_zipfile:
File "C:\Users\Anonymous\AppData\Local\Programs\Python\Python310\lib\site-packages\torch\serialization.py", line 242, in init
super(_open_zipfile_reader, self).init(torch._C.PyTorchFileReader(name_or_buffer))
RuntimeError: PytorchStreamReader failed reading zip archive: failed finding central directory

[lukas-blecher]

Looks like the weights were not completely downloaded.
To fix this please run:

python -m pix2tex.model.checkpoints.get_latest_checkpoint

[aghitshek]

It worked i also install [gui] how do i use it

…

On Thu, 6 Oct, 2022, 5:26 pm Lukas Blecher, ***@***.***> wrote: Looks like the weights were not completely downloaded. To fix this please run: python -m pix2tex.model.checkpoints.get_latest_checkpoint — Reply to this email directly, view it on GitHub <#202 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/A3NOORTSYR4HIYJNG3KEOOLWB2475ANCNFSM6AAAAAAQ4SRM3U> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[lukas-blecher]

by calling latexocr.

If you get a dependency error, maybe you need to install it with the upgrade flag -U

pip install -U pix2tex[gui]

[aghitshek]

Thank you.
does pix2tex convert only one equation at a time?

it will be amazing if it would work like tesseract( Tesseract + pix2tex) :)

[lukas-blecher]

Yes, at the moment only one at a time.
You can get the prediction of multiple images with the CLI, by specifying a folder or using a grob but it will be handled sequentially.
But I'm working on a better solution rn...

[aghitshek]

May i help you in one way or another? :)

…

On Fri, 7 Oct, 2022, 4:46 am Lukas Blecher, ***@***.***> wrote: Yes, at the moment only one at a time. But I'm working on it... — Reply to this email directly, view it on GitHub <#202 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/A3NOORXBZWFQWD2JNMQO453WB5MWJANCNFSM6AAAAAAQ4SRM3U> . You are receiving this because you authored the thread.Message ID: ***@***.***>