Dev Retrospective: Real-time Guidance Embedding Training for Flex Models
ostrisai · x · 2026-08-05
Developer ostris shared a technical anecdote: while working on Contrastive Guidance Loss for MiniMax H3, he realized he had previously used a similar method of continuously training the guidance embedding on the fly for Flex.1 and Flex.2 models, completely forgetting it was still in the codebase.
More from Models
- Reflection AI open-sources Beam under Apache 2.0 with 40+ ecosystem partners — zhyncs42 · 2026-10-16
- Why are sub-15B open models so rare now? Labs seem to have settled on 24-27B minimum — Hot_Example_4456 · 2026-10-16
- User spots model openly generating "ideal responses" to train other models — liminal_bardo · 2026-10-16
- Reflection releases open MoE Beam-1.0: 501B total/23B active, 1M context, 3x fewer reasoning tokens — jacek2023 · 2026-10-16
- Reflection's 501B MoE Beam hits 463 tok/s on 8x GB200 with NVFP4 at day 0 — zhyncs42 · 2026-10-16
- Gemma 4 31B QAT on one 24GB card: 46 tok/s plain, ~80 with draft model on RTX 4090 — Subject_Poet_3861 · 2026-10-16