config.json

{
  "_name_or_path": "/mnt/hwfile/medai/LLMModels/Model/Qwen1.5-7B-Chat",
  "architectures": [
    "MoLoRAQwenForCausalLM"
  ],
  "attention_dropout": 0.0,
  "bos_token_id": 151643,
  "enable_router_loss": false,
  "eos_token_id": 151645,
  "expert_selection": "top_k",
  "hidden_act": "silu",
  "hidden_size": 4096,
  "initializer_range": 0.02,
  "intermediate_size": 11008,
  "lbl_loss_coeff": 0.0,
  "max_position_embeddings": 32768,
  "max_window_layers": 28,
  "model_type": "qwen2",
  "num_attention_heads": 32,
  "num_experts": 8,
  "num_experts_per_token": 2,
  "num_hidden_layers": 32,
  "num_key_value_heads": 32,
  "num_share_experts": 2,
  "orth_loss_weight": 0.0,
  "return_activated_experts_index": false,
  "rms_norm_eps": 1e-06,
  "rope_theta": 1000000.0,
  "router_loss_coeff": 0.0,
  "router_loss_mode": "contrastive",
  "share_expert": true,
  "sliding_window": 32768,
  "tie_word_embeddings": false,
  "torch_dtype": "float16",
  "transformers_version": "4.41.0.dev0",
  "use_cache": true,
  "use_lbl_loss": false,
  "use_rslora": false,
  "use_sliding_window": false,
  "vocab_size": 151936
}