{
  "architectures": [
    "NanoHammerForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "auto_map": {
    "AutoConfig": "NanoHammerForCausalLM.NanoHammerConfig",
    "AutoModelForCausalLM": "NanoHammerForCausalLM.NanoHammerForCausalLM"
  },
  "bos_token_id": 128000,
  "dtype": "bfloat16",
  "eos_token_id": 128009,
  "hidden_act": "silu",
  "hidden_size": 2048,
  "initializer_range": 0.02,
  "intermediate_size": 8192,
  "max_position_embeddings": 131072,
  "mlp_bias": false,
  "model_type": "nanohammer",
  "num_attention_heads": 32,
  "num_hidden_layers": 16,
  "num_key_value_heads": 8,
  "num_state_heads": 16,
  "pad_token_id": 128009,
  "rms_norm_eps": 1e-05,
  "rope_scaling": null,
  "rope_theta": 500000.0,
  "state_hidden_size": 512,
  "tie_word_embeddings": false,
  "transformers_version": "4.57.6",
  "use_cache": true,
  "vocab_size": 128256
}