{
  "architectures": [
    "NemotronForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "bos_token_id": 2,
  "dtype": "bfloat16",
  "eos_token_id": 3,
  "head_dim": 128,
  "hidden_act": "relu2",
  "hidden_size": 3072,
  "initializer_range": 0.0134,
  "intermediate_size": 9216,
  "kv_channels": 128,
  "max_position_embeddings": 4096,
  "mlp_bias": false,
  "model_type": "nemotron",
  "norm_eps": 1e-05,
  "num_attention_heads": 24,
  "num_hidden_layers": 32,
  "num_key_value_heads": 8,
  "partial_rotary_factor": 0.5,
  "quantization_config": {
    "_load_in_4bit": true,
    "_load_in_8bit": false,
    "bnb_4bit_compute_dtype": "bfloat16",
    "bnb_4bit_quant_storage": "bfloat16",
    "bnb_4bit_quant_type": "nf4",
    "bnb_4bit_use_double_quant": true,
    "llm_int8_enable_fp32_cpu_offload": false,
    "llm_int8_has_fp16_weight": false,
    "llm_int8_skip_modules": null,
    "llm_int8_threshold": 6.0,
    "load_in_4bit": true,
    "load_in_8bit": false,
    "quant_method": "bitsandbytes"
  },
  "rope_theta": 10000,
  "tie_word_embeddings": false,
  "transformers_version": "4.57.1",
  "use_cache": false,
  "vocab_size": 256000
}