تحسين نموذج Arabic-GLM-OCR-V2

#1
by sherif1313 - opened

السلام عليكم اخي حسن
هل تستطيع ان تشاركني في تحسين نموذج "جلم 2" بمجموعة البيانات الخاصة بك وان ارد ضبط النموذج اخبرني وساعطيك بيانات التدريب الخاص بي ان كنت تريد .... واني اعمل الان علي نموذج صوتي بحجم 100 ميجا ويعتبر اصغر نموذج صوتي حت الان ان كنت تريد ممكن نتشارك في وكنت عملت به من الصفر فهو نموذج صغير الحجم للكتب الدينية وان شاء الله عند الانتهاء منه سوف انزله

sure eng.hassan
this is the dataset for the OCR https://huggingface.co/datasets/U4RASD/omar-al-saleh-manuscripts-segments , https://huggingface.co/datasets/U4RASD/omar-al-saleh-manuscripts-full
for the Speach model i would love to work with u

eng.hassan, you're very late... The model is ready for training now. If you have time, could you help me with it? It's a very powerful model and will be stronger than all the other trading models, but it's difficult for me to continue working on it alone. Perhaps we can share data together to improve it. It needs a lot of time and practice.
I'm currently working on a voice_design, which is based on the original template, and I'll add it within a few days.

Sign up or log in to comment