alwadifa's picture

alwadifa

vnrom

·

https://alouadifa.ma/

AI & ML interests

Alouadifa.ma is a leading Moroccan digital platform dedicated job recruitment, professional training, business networking, etc.]. Designed to bridge gaps between employers and talent, students and institutions, the platform offers innovative solutions tailored to Morocco’s dynamic market. With a user-friendly interface, advanced search tools, and a commitment to excellence, Alouadifa.ma empowers users to find dream jobs, upskill, grow their businesses. Whether you’re a, the platform provides trusted, localized resources to help you thrive in today’s competitive landscape.

Recent Activity

replied to anakin87's post 9 days ago

𝗜 𝘁𝗿𝗮𝗶𝗻𝗲𝗱 𝗮 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗠𝗼𝗱𝗲𝗹 𝘁𝗼 𝘀𝗰𝗵𝗲𝗱𝘂𝗹𝗲 𝗲𝘃𝗲𝗻𝘁𝘀 𝘄𝗶𝘁𝗵 𝗚𝗥𝗣𝗢! 👑 🗓️ ✍️ Blog post: https://huggingface.co/blog/anakin87/qwen-scheduler-grpo I experimented with GRPO lately. I am fascinated by models learning from prompts and rewards - no example answers needed like in Supervised Fine-Tuning. After the DeepSeek boom, everyone is trying GRPO with GSM8K or the Countdown Game... I wanted a different challenge, like 𝘁𝗲𝗮𝗰𝗵𝗶𝗻𝗴 𝗮 𝗺𝗼𝗱𝗲𝗹 𝘁𝗼 𝗰𝗿𝗲𝗮𝘁𝗲 𝗮 𝘀𝗰𝗵𝗲𝗱𝘂𝗹𝗲 𝗳𝗿𝗼𝗺 𝗮 𝗹𝗶𝘀𝘁 𝗼𝗳 𝗲𝘃𝗲𝗻𝘁𝘀 𝗮𝗻𝗱 𝗽𝗿𝗶𝗼𝗿𝗶𝘁𝗶𝗲𝘀. Choosing an original problem forced me to: 🤔 Think about the problem setting 🧬 Generate data 🤏 Choose the right base model 🏆 Design reward functions (and experiencing reward hacking) 🔄 Run multiple rounds of training, hoping that my model would learn something. A fun and rewarding 😄 experience. I learned a lot of things, that I want to share with you. 👇 ✍️ Blog post: https://huggingface.co/blog/anakin87/qwen-scheduler-grpo 💻 Code: https://github.com/anakin87/qwen-scheduler-grpo 🤗 Hugging Face collection (dataset and model): https://huggingface.co/collections/anakin87/qwen-scheduler-grpo-680bcc583e817390525a8837

View all activity

Organizations

None yet

vnrom's activity

replied to anakin87's post 9 days ago

Love how you stepped off the beaten path and tackled a more practical, structured challenge training a model to build schedules is no small feat! GRPO really shines in these setups where reward design becomes the art. Your experience captures the real spirit of experimentation: data crafting, reward tuning, and watching models learn from scratch. Curious to hear how your reward functions evolved especially how you handled reward hacking. Great work!