Tag
The article questions whether top AI models will become limited to governments and a few companies, leaving consumers with less capable models.
NVIDIA introduces ModelExpress, a tool that accelerates distribution of model artifacts across GPU clusters by using GPU-to-GPU RDMA transfers and optimized streaming from storage, reducing model startup times from 8 minutes to under 2 minutes for large models like DeepSeek-V4 Pro.
NVIDIA introduces ModelExpress, a solution for rapidly distributing AI model artifacts, as described in their technical blog.
A repo and site for sharing .torrent files for open models, using Hugging Face as a web seed fallback for peerless downloads.
Noema Atlas is a free and open-source peer-to-peer desktop app for decentralized distribution of LLM model weights, using content-addressed verification and Iroh for direct machine-to-machine transfers, with Hugging Face as a fallback.