config.json

{
  "architectures": [
    "IQuestLoopCoderForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "auto_map": {
    "AutoConfig": "configuration_iquestloopcoder.IQuestLoopCoderConfig",
    "AutoModel": "modeling_iquestloopcoder.IQuestLoopCoderModel",
    "AutoModelForCausalLM": "modeling_iquestloopcoder.IQuestLoopCoderForCausalLM"
  },
  "bos_token_id": 1,
  "dtype": "bfloat16",
  "eos_token_id": [
    75864,
    2,
    75869
  ],
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 5120,
  "initializer_range": 0.02,
  "intermediate_size": 27648,
  "loop_num": 2,
  "loop_window_size": 64,
  "max_position_embeddings": 16384,
  "mlp_bias": false,
  "model_type": "iquestloopcoder",
  "num_attention_heads": 40,
  "num_hidden_layers": 80,
  "num_key_value_heads": 8,
  "rms_norm_eps": 1e-05,
  "rope_scaling": null,
  "rope_theta": 500000,
  "tie_word_embeddings": false,
  "transformers_version": "4.56.0",
  "use_cache": true,
  "vocab_size": 76800,
  "quantization_config": {
    "config_groups": {
      "group_0": {
        "input_activations": {
          "dynamic": false,
          "num_bits": 4,
          "type": "float",
          "group_size": 16
        },
        "weights": {
          "dynamic": false,
          "num_bits": 4,
          "type": "float",
          "group_size": 16
        },
        "targets": [
          "Linear"
        ]
      }
    },
    "ignore": [
      "lm_head"
    ],
    "quant_algo": "NVFP4",
    "producer": {
      "name": "modelopt",
      "version": "0.41.0"
    },
    "quant_method": "modelopt"
  }
}