framework

Hugging Face Transformers

Listed in two builds as the mechanism for loading and running the Qwen language model.

Current clusters

build1 distinct publisher

NeoMME runs document retrieval through one from-scratch encoder at 51 pages a second

The 260M NeoMME sends raw image patches and text down the same bidirectional path, with no vision tower and no causal decoder, and reports a 6 kB late-interaction index per page. The conditions behind both numbers are checkable.

Publishers:huggingface.co

Reality

Evidence45
Adoption18
Hype gap+20
Incentives75

build1 distinct publisher

213 seconds per agent step evicts hybrid RAG from the local CPU

The FAISS-plus-BM25 retrieval in this writeup does address vocabulary mismatch, but the agent loop around it ran at over 213 seconds a step on CPU, and that figure decided the deployment, not the retrieval design.

Publishers:dev.to

Reality

Evidence34
Adoption14