config.json

{
  "architectures": [
    "HATForCausalLM"
  ],
  "auto_map": {
    "AutoConfig": "config.HATArchitectureConfig",
    "AutoModelForCausalLM": "model.HATForCausalLM"
  },
  "backbone_config": {
    "hidden_size": 4096,
    "intermediate_size": 14336,
    "is_neox_style": true,
    "key_query_norm": true,
    "key_query_norm_per_head": true,
    "max_position_embeddings": 3500,
    "mlp_bias": false,
    "num_attention_heads": 32,
    "num_hidden_layers": 32,
    "num_key_value_heads": 8,
    "rms_norm_eps": 1e-05,
    "rope_scaling": {
      "rope_type": "default"
    },
    "rope_theta": 500000,
    "sliding_window": null,
    "transformers_version": null,
    "use_cache": true,
    "vocab_size": 0
  },
  "decoder_config": {
    "cross_attention_config": {
      "attention_num_kv_heads": 8,
      "hidden_size": 1024,
      "hidden_size_kv": 4096,
      "hidden_size_q": 1024,
      "key_query_norm": true,
      "key_query_norm_per_head": true,
      "num_attention_heads": 8,
      "word_window_size": 1
    },
    "cross_attn_every_layer": true,
    "hidden_size": 1024,
    "intermediate_size": 2816,
    "is_neox_style": true,
    "key_query_norm": true,
    "key_query_norm_per_head": true,
    "max_position_embeddings": 28000,
    "mlp_bias": false,
    "num_attention_heads": 8,
    "num_hidden_layers": 4,
    "num_key_value_heads": 8,
    "rms_norm_eps": 1e-05,
    "rope_scaling": {
      "rope_type": "default"
    },
    "rope_theta": 100000,
    "sliding_window": 768,
    "transformers_version": null,
    "use_cache": true,
    "vocab_size": 256
  },
  "encoder_config": {
    "cross_attention_config": {
      "attention_num_kv_heads": 32,
      "hidden_size": 4096,
      "hidden_size_kv": 1024,
      "hidden_size_q": 4096,
      "key_query_norm": true,
      "key_query_norm_per_head": true,
      "num_attention_heads": 32,
      "word_window_size": 1
    },
    "hidden_size": 1024,
    "intermediate_size": 2816,
    "is_neox_style": true,
    "key_query_norm": true,
    "key_query_norm_per_head": true,
    "max_position_embeddings": 28000,
    "mlp_bias": false,
    "num_attention_heads": 8,
    "num_hidden_layers": 6,
    "num_key_value_heads": 8,
    "rms_norm_eps": 1e-05,
    "rope_scaling": {
      "rope_type": "default"
    },
    "rope_theta": 100000,
    "sliding_window": 768,
    "transformers_version": null,
    "use_cache": true,
    "vocab_size": 256
  },
  "max_position_embeddings": 28000,
  "max_word_size": 100,
  "model_type": "hierarchical_autoregressive_transformer",
  "sliding_window": 768,
  "special_token_dict": {},
  "torch_dtype": "bfloat16",
  "transformers_version": "4.46.3"
}