{
  "add_qk_norm": true,
  "architectures": [
    "TPULlamaForCausalLM"
  ],
  "attention_bias": true,
  "attention_dropout": 0.0,
  "auto_map": {
    "FlaxAutoModelForCausalLM": "modelling_flax_tpu_llama.FlaxTPULlamaForCausalLM"
  },
  "bos_token_id": 151643,
  "eos_token_id": 151643,
  "expand_input_ids": false,
  "expand_input_ids_dict": null,
  "expand_input_ids_maxlen": null,
  "expand_input_ids_vocab_size": null,
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 2560,
  "initializer_range": 0.02,
  "intermediate_size": 9728,
  "max_length": 8192,
  "max_position_embeddings": 32768,
  "max_window_layers": 36,
  "mlp_bias": false,
  "model_type": "tpu_llama",
  "num_attention_heads": 32,
  "num_hidden_layers": 36,
  "num_key_value_heads": 8,
  "pretraining_tp": 1,
  "rms_norm_eps": 1e-06,
  "rope_scaling": null,
  "rope_theta": 1000000,
  "skip_out_norm": false,
  "sliding_window": null,
  "tie_word_embeddings": true,
  "torch_dtype": "float32",
  "transformers_version": "4.52.3",
  "use_cache": true,
  "use_sliding_window": false,
  "vocab_size": 151936
}