@TrainLoop_ai
The best logistics model isn't Astra or Fable.
It's a small model, trained on lessons from real shipments. It beat every frontier model Cartage tested on its freight benchmark, while running 20x cheaper than the frontier.
Cartage-1 powers Wilson, unlocking significantly more complex logistics workflows than consumer models.View quoted post on X
We partnered with @nvidia, @nollahealth, and @nebiusai to post train Nemotron 3.5 VL Super far past the frontier on dermatology diagnosis.
The model set new standards in our internal measures of trainability, and with Nolla it's powering the future of AI in medical care.View quoted post on X
can we train a model in single RL step?
During recent experiments, @Logangrasby found that a single step of OAPL increased model performance from ~0 to 48% on a clinical reasoning and prediction task.
Turns out, data staleness might matter less than we think. with @pathos :View quoted post on X
We trained LoRA adapters of different ranks to understand training dynamics, finding that adapters for GSM8k live in a surprisingly vast, low-rank solution space.
This hints that some model skills are easy to learn, and training is more forgiving than we think. @hasith_v 1/6 🧵View quoted post on X
