config.json

{
  "model_type": "nanogpt",
  "architectures": [
    "NanoGPTForCausalLM"
  ],
  "vocab_size": 16388,
  "hidden_size": 576,
  "num_hidden_layers": 14,
  "num_attention_heads": 9,
  "num_key_value_heads": 1,
  "head_dim": 64,
  "intermediate_size": 1664,
  "max_position_embeddings": 2048,
  "rope_theta": 10000.0,
  "rms_norm_eps": 1e-05,
  "tie_word_embeddings": true,
  "torch_dtype": "float32",
  "auto_map": {
    "AutoConfig": "configuration_nanogpt.NanoGPTConfig",
    "AutoModelForCausalLM": "modeling_nanogpt.NanoGPTForCausalLM"
  }
}