Brand is building a solution for shared reinforcement learning workflows.
Harbor is really great. I like the design and it's well polished for doing evals. It would be great to use the same rollout utility for everything (RL / eval / new tasks definitions).
Rollouts for eval, rollouts for RL, rollouts for GEPA, rollouts for prod, rollouts for trajectory analysis, rollouts for SFT data gen, rollouts rollouts rollouts
Users are excited about Harbor unifying rollouts across AI evals, RL, and optimization loops because the announcement prompts sharing of relevant tools and insights.
No Digg Deeper questions have been answered for this story yet.
@vwxyzjn i wish it was a better fit for tasks that arent coding agent in a sandbox shaped. at least that was my read when i evaluated using it to train a search agent
@vwxyzjn Oh boy do I have something for you
Harbor is really great. I like the design and it's well polished for doing evals. It would be great to use the same rollout utility for everything (RL / eval / new tasks definitions).
@andersonbcdefg why can't a search based task not a good fit for harbor's task definition?
@andersonbcdefg @vwxyzjn https://github.com/hud-evals/hud-python for tasks of any shape btw