MindGPT-4ov: An Enhanced MLLM via a Multi-Stage Post-Training Paradigm Paper • 2512.02895 • Published Dec 2, 2025 • 5
MindWatcher: Toward Smarter Multimodal Tool-Integrated Reasoning Paper • 2512.23412 • Published 14 days ago • 36
MindGPT-4ov Collection MindGPT-4ov is a MLLM that introduces innovative data generation, SFT, and RL methods, achieving SOTA performance and efficiency. • 1 item • Updated Dec 8, 2025 • 2
Running 3.64k The Ultra-Scale Playbook 🌌 3.64k The ultimate guide to training LLM on large GPU Clusters