HF: transformers backend now matches native vLLM speed, no porting needed
ariG23498 · x · 2026-09-28
Hugging Face announced the vLLM transformers modeling backend now meets or beats hand-written native vLLM implementations across Qwen3 4B dense, 32B tensor-parallel, and 235B FP8 MoE setups—so model authors integrated into transformers get fast vLLM inference for free, no porting required.
More from Infra
- Developer says local AI is shifting from nice-to-have to infrastructure: control beats privacy — Aiden_Tech_Ai · 2026-09-28
- Meta open-sources Component Benchmark, a hierarchical profiler for TB-scale recommender models — _reachsumit · 2026-09-28
- apple-llm: Node/Python wrapper for the free local LLM built into Apple Silicon Macs — light_2earth · 2026-09-28
- Running 8 watercooled GPUs for local AI: one user's case for watercooling over air cooling — HanchungLee · 2026-09-28
- Renting their cluster's compute would have cost over $1 billion on a 5-year deal — ericzelikman · 2026-09-28
- Self-Hosted K3 Cluster Lets Autonomous Infra Research Burn Tokens Without Worry — Xianbao_QIAN · 2026-09-28