{
  "architectures": [
    "MiniMindLM"
  ],
  "aux_loss_alpha": 0.1,
  "dim": 512,
  "dropout": 0.0,
  "flash_attn": true,
  "hidden_dim": 1408,
  "max_seq_len": 8192,
  "model_type": "minimind",
  "multiple_of": 64,
  "n_heads": 8,
  "n_kv_heads": 2,
  "n_layers": 8,
  "n_routed_experts": 4,
  "n_shared_experts": true,
  "norm_eps": 1e-05,
  "norm_topk_prob": true,
  "num_experts_per_tok": 2,
  "rope_theta": 1000000.0,
  "scoring_func": "softmax",
  "seq_aux": true,
  "torch_dtype": "float32",
  "transformers_version": "4.50.2",
  "use_moe": false,
  "vocab_size": 6400,
  "task_type": "text-generation",
  "auto_map": {
    "AutoModelForCausalLM": ["model.model.MiniMindLM", null]
  },
  "generation_config": {
    "max_length": 2048,
    "temperature": 0.7,
    "top_p": 0.9,
    "do_sample": true,
    "pad_token_id": 0,
    "eos_token_id": 2,
    "bos_token_id": 1
  }
}