erikbern / deep-fonts Goto Github PK

I'm trying to download the fonts.hdf5 file but it is saying I need access. Would you be able to adjust the permissions please?

Thanks!

Extremely interesting project, can you provide a readme or getting started guide?

i am so excited to learn about your work about generation of characters. i am a college student,and i am doing this relative work. so i want to learn your finished work by coding .but i can not get the training data. could you share your training data set or website? thank you!!

Hey, could you help me with the error below?
UnicodeDecodeError: 'ascii' codec can't decode byte 0x93 in position 0: ordinal not in range(128)

Maybe the model.pickle.gz file is corrupted.

Also, What python version did you use to run this code?

Thank you.

What is the licensing intent for this project? GPLv3? :)

I'm new to python language, I'm trying to create my own dataset so I can use it for training.
I have all the .ttf files inside the folder "scraper/fonts". But when I run create_dataset.py, I get the following error:

python create_dataset.py
Traceback (most recent call last):
File "create_dataset.py", line 78, in
for fn in get_ttfs(d=sys.argv[1]):
IndexError: list index out of range

Is there a easy solution to that?
Thank you

maybe this is on purpose, but this url - https://s3.amazonaws.com/erikbern/fonts.hdf5 - no longer works.

Hi Erik,

would it be possible to get the original .ttf files (maybe alongside the scrapy scripts)? I am planning to use the fonts for research and need to render fonts in different (higher than 64x64) resolutions.

Thanks!

I read with interest your article and looked through the script. Thank you for this great contribution! Since it takes several weeks to train the model is there a way to get the trained model ?

Is there any relationship between this repo and zi2zi? And what would be the difference in models?

@erikbern How can I use the trained model to detect the font, for example:
I have a scanned image and want to detect the used font name in that image, anyway?
waiting for your reply