config.json

{
  "adapter_alpha": 1.0,
  "architectures": [
    "AttnQwenForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "dtype": "float32",
  "eos_token_id": 151645,
  "from_pretrained_init": false,
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 1024,
  "initializer_range": 0.02,
  "intermediate_size": 3072,
  "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"
  ],
  "low_rank": null,
  "lowrank_link": 16,
  "max_position_embeddings": 32768,
  "max_window_layers": 28,
  "mod_k_proj": true,
  "mod_keys": true,
  "mod_output": true,
  "mod_queries": true,
  "mod_v_proj": true,
  "mod_vals": false,
  "modattn_layers": "all",
  "modattn_type": "decoder",
  "model_type": "attnqwen3",
  "num_attention_heads": 16,
  "num_hidden_layers": 28,
  "num_key_value_heads": 8,
  "num_modattn_kvheads": 8,
  "pad_token_id": 151643,
  "rms_norm_eps": 1e-06,
  "rope_scaling": null,
  "rope_theta": 10000.0,
  "sliding_window": null,
  "sssm_hidden_dim": 16,
  "sssm_low_rank": 8,
  "tie_word_embeddings": false,
  "transformers_version": "4.57.3",
  "use_S0": true,
  "use_cache": false,
  "use_phi": true,
  "use_sliding_window": false,
  "use_sssd": true,
  "use_val_rope": true,
  "vocab_size": 152000
}