Mystery of falling HF downloads: local GGUF/AWQ/ONNX model downloads are surging
xeophon · x · 2026-10-14
Nathan Lambert had tracked a 30% ongoing decline in HuggingFace downloads for top open LLMs. He now reports a methodology gap: local formats (GGUF, AWQ, ONNX) weren't counted — once a rounding error, they're now surging. The interconnects pipeline has been updated to track them.
Related event: HuggingFace Downloads Plateau as Local Model Deployments Surge(4 posts)→
More from Infra
- Same model, 9x cost gap: Artificial Analysis publishes cache hit rates per provider — ArtificialAnlys · 2026-10-14
- Apple's Minions turns idle Mac GPUs into a pooled cloud for ML inference — MikePFrank · 2026-10-14
- AnythingLLM ditches Ollama for llama.cpp, with up to 2x faster inference — solyarisoftware · 2026-10-14
- LlamaForge: a browser panel for upstream llama.cpp with local agents that must quote their sources — Sleepybear2611 · 2026-10-14
- CoreWeave's GPU fleet is mostly older boxes, raising questions about depreciation math — ivan_bezdomny · 2026-10-14
- NVMe reads the 7GB checkpoint in 2s, so why does ComfyUI take 30s to load? — rguerraf · 2026-10-14