FreeAI

community
Activity Feed

AI & ML interests

Post-Training & Agent & Self-Distillation & Pre-Training

Recent Activity

Wuyuhang11Β  updated a Space 5 days ago
freeai-org/README
Wuyuhang11Β  updated a model 6 days ago
freeai-org/Scalpel-VL-1.8B
Wuyuhang11Β  published a model 6 days ago
freeai-org/Scalpel-VL-1.8B
View all activity

Organization Card

The official organization of FreeAI

πŸ’ Introduction: FreeAI is an open research & engineering organization building practical, interesting, deployable AI. We focus on pre-training, post-training, agent systems and domain-specific AI applications, sharing open-source models, engineering practices and research ideas to drive AI from experimental demos to real-world deployment.

✨ Join Us: We welcome researchers, developers, and students who are interested in practical AI systems, and domain-specialized models. If you are interested in FreeAI or would like to collaborate, please contact us.

πŸ“• [2026-8-26 Scalpel-VL-1.6B-Animal is Opened] NEW!!!β€”β€”Scalpel-VL-1.6B-Animal: is a lightweight multimodal model for structured animal behavior recognition. It outputs schema-constrained JSON for feline posture, action, while achieving around 30-40% faster inference than models of comparable scale.

πŸ“• [2026-8-31 Scalpel-VL-1.7B is Opened] NEW!!!β€”β€”Scalpel-VL-1.7B: is a lightweight multimodal model trained by ScalpelBench.

πŸ“• [2026-8-31 ScalpelBench is Opened] NEW!!!β€”β€”ScalpelBench: A 0.1B‑scale dataset with 300K samples spanning English, Chinese, mathematics and code is adopted to retain the base‑model capabilities during pruning.