config.json

{
  "architectures": [
    "VerMindForCausalLM"
  ],
  "auto_map": {
    "AutoConfig": "configuration_vermind.VerMindConfig",
    "AutoModelForCausalLM": "modeling_vermind.VerMindForCausalLM"
  },
  "aux_loss_alpha": 0.01,
  "bos_token_id": 1,
  "dropout": 0.0,
  "dtype": "float32",
  "eos_token_id": 2,
  "flash_attn": true,
  "hidden_act": "silu",
  "hidden_size": 768,
  "inference_rope_scaling": false,
  "intermediate_size": 2048,
  "max_position_embeddings": 32768,
  "model_type": "vermind",
  "n_routed_experts": 4,
  "n_shared_experts": 1,
  "norm_topk_prob": true,
  "num_attention_heads": 8,
  "num_experts_per_tok": 2,
  "num_hidden_layers": 16,
  "num_key_value_heads": 2,
  "rms_norm_eps": 1e-05,
  "rope_scaling": null,
  "rope_theta": 1000000.0,
  "scoring_func": "softmax",
  "seq_aux": true,
  "transformers_version": "4.57.6",
  "use_moe": false,
  "vocab_size": 6400
}