{
 "setting": "KD only",
 "alpha": 1.0,
 "order": "code_math",
 "seed": 0,
 "exp": 2,
 "student": "Qwen/Qwen2.5-1.5B-Instruct",
 "gpu_student": "NVIDIA RTX 6000 Ada Generation",
 "gpu_teachers": "NVIDIA RTX A6000",
 "load_h": 0.003855370150672065,
 "stages": {
  "stage1_code": {
   "domain": "code",
   "steps": 922,
   "train_h": 2.9412914867533577,
   "val_start": {
    "step": 0,
    "math_kl": 0.15902198004314483,
    "math_ce": 0.6553564874444496,
    "code_kl": 0.14797331712898912,
    "code_ce": 0.3033394576202258,
    "math_hard_kl": 0.24902600898986357,
    "math_hard_ce": 0.6769691187959327
   },
   "val_end": {
    "step": 922,
    "math_kl": 0.16046075451441033,
    "math_ce": 0.6183213579269871,
    "code_kl": 0.07800855337397358,
    "code_ce": 0.31004403168757744,
    "math_hard_kl": 0.24882929720080427,
    "math_hard_ce": 0.6718085009279865
   },
   "peak_gb_student": 33.80862855911255,
   "peak_gb_teachers": 30.005428314208984
  },
  "stage2_math": {
   "domain": "math",
   "steps": 896,
   "train_h": 2.814592224955559,
   "val_start": {
    "step": 0,
    "math_kl": 0.16056643818369662,
    "math_ce": 0.6179343579401348,
    "code_kl": 0.07794113298743313,
    "code_ce": 0.309922693351798,
    "math_hard_kl": 0.24884287280595874,
    "math_hard_ce": 0.6717120443366895
   },
   "val_end": {
    "step": 896,
    "math_kl": 0.0759878009343543,
    "math_ce": 0.8331185948595244,
    "code_kl": 0.10406180013954185,
    "code_ce": 0.3197136727378907,
    "math_hard_kl": 0.08557213517622467,
    "math_hard_ce": 0.8143728766945073
   },
   "peak_gb_student": 36.68444871902466,
   "peak_gb_teachers": 30.639182090759277,
   "bench_h": 0.5482511294550366,
   "gpu_bench": "NVIDIA RTX 6000 Ada Generation"
  }
 },
 "resumed": {
  "from": "stage1_code",
  "gpu_student": "NVIDIA RTX 6000 Ada Generation",
  "gpu_teachers": "NVIDIA RTX A6000",
  "load_h": 0.008158521188629999
 },
 "train_h": 5.7558837117089166,
 "done": true
}