Hugging Face transformers now exports and quantizes every model to any backend, drop-in
huggingface · x · 2026-11-14
Hugging Face's Ilys Moutawwakil announced a major transformers update: you can now export, quantize, and run every transformers model end-to-end on any backend.
The key selling point is drop-in replacement — identical inputs produce identical outputs before and after export/quantization, so developers can move models to different inference backends without changing existing code. Docs and release notes are included.
More from Infra
- Investor Reveals Dual NBIS-AKAM Position: Betting Both Ends of the Compute Stack — pdamodaran · 2026-11-14
- Nina Schick: the first computing cycle where demand isn't capped by humans — NinaDSchick · 2026-11-14
- MiniMax H3 Image's ~50GB stack runs on 8GB VRAM at 57s/image — how is there no OOM? — wreck_of_u · 2026-11-14
- AMD raises guidance: revenue set to grow 80%+ next year as data center AI revenue more than doubles — Beth_Kindig · 2026-11-14
- Running Qwen3.8 27B on a 2018 Jetson AGX Xavier: 106 tok/s prefill, power-brick fixes and all — cortexist · 2026-11-14
- Microsoft Was Down 18 Hours — Its SLA Owes You Up to a Free Month — justin_hart · 2026-11-14