How Molmo2Fish was trained: sonar fine-tuning plus error-correction examples
allen_ai · x · 2026-10-14
Ai2 shares training details: Kai Van Brunt fine-tuned Molmo 2 on sonar video, then trained it on examples of tracking errors, instructions, and corrections. In tests, Molmo2Fish fixed some mistakes after guidance — e.g., when two salmon had been assigned the same track. Content overlaps the main thread post; see its summary for full context.
Related event: MIT Researchers Build Molmo2Fish to Track Salmon in Sonar Video(4 posts)→
More from Research
- Spork post-training recipe lifts chess puzzle solve rate 100% over best GRPO setup — aidenybai · 2026-10-14
- Dyna-2 Deep Dive: World-Action Models Trained on 1M Hours of Human Video, 99.4% 24-Hour Autonomy — Latent Space · 2026-10-14
- Procedural Core: a 1M-param abstract model boosts 85M-param ViT ImageNet accuracy by 2.2 points — kastnerkyle · 2026-10-14
- Nature Will unveils Natura-1: 50B-parameter MoE robot foundation model trained on 1M hours of egocentric human video — rohanpaul_ai · 2026-10-14
- PR-OPD aligns hidden states for agentic RL, up to +14.0 WebShop accuracy over GRPO — Muyang Li · 2026-10-14
- Do reasoning traces actually matter? A three-stage causal study on distilled DeepSeek R1 models — burkov · 2026-10-14