{
  "model_id": "meta-llama/Llama-3.2-3B",
  "task": "rte",
  "glue_config": "rte",
  "bitwidth": "nf4",
  "seed": 43,
  "lora": {
    "r": 16,
    "alpha": 32,
    "dropout": 0.0,
    "bias": "none",
    "target_modules": [
      "q_proj",
      "k_proj",
      "v_proj",
      "o_proj"
    ]
  },
  "hyperparams": {
    "num_train_epochs": 3,
    "learning_rate": 0.0002,
    "max_length": 256,
    "per_device_train_batch_size": 4,
    "per_device_eval_batch_size": 8,
    "gradient_accumulation_steps": 16,
    "warmup_ratio": 0.03,
    "lr_scheduler_type": "cosine",
    "gradient_checkpointing": true
  },
  "lora_init_source": "shared:lora_init_3B_seed43.pt:224tensors"
}