{
 "base_model": "Qwen/Qwen3.5-2B",
 "lora_rank": 32,
 "lora_init_seed": null,
 "data_seed": 0,
 "optimizer": {
  "type": "adamw",
  "learning_rate": 2e-05,
  "beta1": 0.9,
  "beta2": 0.95,
  "eps": 1e-08,
  "weight_decay": 0.0,
  "grad_clip_norm": 0.0
 },
 "renderer": "qwen3_5_disable_thinking",
 "env": "tinker_cookbook 0.5.9 math_rl Gsm8kDataset/MathEnv, standard fewshot prefix, format_coef 0.1",
 "groups_per_batch": 32,
 "group_size": 8,
 "temperature": 1.0,
 "max_tokens": 512,
 "advantages": "group-mean centered (cookbook compute_advantages)",
 "substeps_per_batch": 2,
 "minibatch_split": "groups 0-15 / 16-31",
 "planned_steps": 50,
 "executed_pilot_steps": 5,
 "full_training_executed": false,
 "grpo": {
  "eps": 0.2
 },
 "dppo": {
  "variant": "Binary-TV",
  "delta": 0.15,
  "clip_ratio_c": 10.0
 },
 "loss_normalization": "token sum",
 "endpoint": "shared Spindle (TINKER_BASE_URL)",
 "init_state_path": "tinker://21e32976f18f971463410adff4e364e4/weights/rx19badd842c_init",
 "eval": {
  "temperature": 0.0,
  "max_tokens": 512,
  "stop": [
   248046
  ],
  "num_samples": 1
 }
}