lodestone-rock/RamTorch
Trains models exceeding GPU memory
View on index · View in 3D Map
Related repos
- memory operating system for managing AI model memory across tasks and sessions
- Run 70B LLMs on a 4GB GPU with layer-wise inference and memory optimization, quantization optional
- Persistent memory system for Claude Code sessions
- Hardware-accelerated GPU terminal emulator
- open-source memory engine for LLMs and AI agents
- High-performance OpenGL C++17 3D graphics engine
- Builds knowledge through conversations with LLMs
- Lifelong memory for LLM agents
- GPU-optimized library for training massive transformer models at scale
- Containers for LLMs on AMD Radeon GPUs
