{
  "model_type": "ramo",
  "architectures": [
    "RamoForCausalLM"
  ],
  "auto_map": {
    "AutoConfig": "configuration_ramo.RamoConfig",
    "AutoModelForCausalLM": "modeling_ramo.RamoForCausalLM"
  },
  "vocab_size": 46957,
  "hidden_size": 768,
  "intermediate_size": 2048,
  "num_hidden_layers": 24,
  "num_attention_heads": 12,
  "num_key_value_heads": 4,
  "max_position_embeddings": 2048,
  "original_max_position_embeddings": 512,
  "attention_dropout": 0.0,
  "attention_implementation": "auto",
  "rope_config": {
    "rope_type": "yarn",
    "rope_theta": 1000000.0,
    "factor": 4.0,
    "partial_rotary_factor": 1.0,
    "beta_fast": 32,
    "beta_slow": 1,
    "mscale": null,
    "mscale_all_dim": null,
    "attention_factor": null
  },
  "moe_config": {
    "intermediate_size": 1024,
    "n_dense_layer": 1,
    "num_experts_per_tok": 2,
    "n_shared_experts": 1,
    "n_routed_experts": 8,
    "routed_scaling_factor": 1.0,
    "seq_aux": true,
    "norm_topk_prob": true
  },
  "use_qk_norm": true,
  "bos_token_id": 1,
  "eos_token_id": 2,
  "pad_token_id": 46945,
  "unk_token_id": 46946,
  "tie_word_embeddings": false
}