config.json

{
  "architectures": [
    "Qwen2ForCausalLM"
  ],
  "attention_dropout": 0.0,
  "aux_loss_type": "mseloss",
  "bos_token_id": 151643,
  "check_data_cls_loss": false,
  "del_layers": [],
  "dtype": "bfloat16",
  "eos_token_id": 151645,
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 2048,
  "initializer_range": 0.02,
  "intermediate_size": 18944,
  "kl_temperature": 10.0,
  "layer_types": [
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention",
    "full_attention"
  ],
  "max_position_embeddings": 32768,
  "max_window_layers": 28,
  "model_type": "qwen2",
  "num_attention_heads": 28,
  "num_hidden_layers": 28,
  "num_key_value_heads": 4,
  "quantize": true,
  "rms_norm_eps": 1e-06,
  "rope_scaling": null,
  "rope_theta": 1000000.0,
  "sliding_window": null,
  "student_attn_from_scratch": false,
  "target_hidden_size": 2048,
  "target_rms_norm_eps": 1e-05,
  "tie_word_emb_proj": false,
  "tie_word_embeddings": false,
  "transformers_version": "4.56.2",
  "use_all_attn": 1,
  "use_attn_map": false,
  "use_aux_loss": true,
  "use_cache": true,
  "use_in_out_mlp": false,
  "use_logits_loss": true,
  "use_ntp_loss": true,
  "use_sliding_window": false,
  "use_std_like_attn": false,
  "vocab_size": 152064
}