rt1/training_recipe.json

{
  "seed": 20260915,
  "sft_steps": 50000,
  "sft_global_batch": 64,
  "core_lr": 1e-05,
  "adapter_lr": 0.0001,
  "warmup_steps": 500,
  "rl_steps": 500,
  "rl_lr": 1e-06,
  "kl_coef": 0.001,
  "clip": 0.2,
  "horizon": 10,
  "groups": 4,
  "beams_per_group": 2,
  "diversity_penalty": 0.1,
  "max_text_tokens": 64,
  "text_layers": [
    3,
    7,
    11
  ],
  "text_inner": 256
}