{
  "dataset": "smoltalk",
  "model_path": "/data/checkpoints/final",
  "training_args": {
    "model_path": "/data/checkpoints/final",
    "dataset": "smoltalk",
    "seq_len": 2048,
    "per_device_batch_size": 4,
    "gradient_accumulation_steps": 4,
    "learning_rate": 2e-05,
    "min_lr_ratio": 0.0,
    "warmup_ratio": 0.05,
    "weight_decay": 0.01,
    "max_grad_norm": 1.0,
    "adam_beta1": 0.9,
    "adam_beta2": 0.999,
    "adam_eps": 1e-08,
    "num_epochs": 1,
    "log_interval": 10,
    "save_interval": 500,
    "seed": 42,
    "bf16": true,
    "output_dir": "/data/checkpoints",
    "tensorboard_dir": "/data/sft-logs",
    "hf_repo": "soyrsoyr/erebus-v2-1.5b-instruct",
    "max_samples": 0
  },
  "total_steps": 16245,
  "num_examples": 1039630,
  "param_count": 1720028160,
  "gpu_count": 4
}