Ornith-1.5 releases DFlash draft checkpoints for 9B/397B/35B-A3B speculative decoding
jacek2023 · reddit · 2026-09-29
ornith-ai released three DFlash draft-model checkpoints on Hugging Face, providing speculative decoding acceleration for Ornith-1.5-9B, Ornith-1.5-397B, and Ornith-1.5-35B-A3B. The DFlash checkpoints are not standalone language models and must run alongside the target models in a DFlash-enabled inference setup; serving configurations are included.
More from Infra
- Shaw mocks data center opponents: hating compute while using the internet is incoherent — zealcaiden · 2026-09-29
- Modeling 1B agent VMs by 2030: what personal AI agents mean for CPU demand — AccBalanced · 2026-09-29
- Tessera: retrieval-driven KV cache reuse cuts RAG serving TTFT by up to 3.6x — _reachsumit · 2026-09-29
- Venice's tokenized inference-credit model is being copied — and oversupply looms — 0xJeff · 2026-09-29
- Model Casting: Mid-Training Recipe Sparsifies FFN Activations for Fewer FLOPs — francoisfleuret · 2026-09-29
- antirez: give companies a fleet of DGX Sparks and Mac Ultras to survive API quota limits — antirez · 2026-09-29