config.json

{
  "architectures": [
    "EmbformerForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "auto_map": {
    "AutoConfig": "configuration_embformer.EmbformerConfig",
    "AutoModelForCausalLM": "modeling_embformer.EmbformerForCausalLM"
  },
  "bos_token_id": 1,
  "eos_token_id": 2,
  "head_dim": 96,
  "hidden_act": "silu",
  "hidden_size": 768,
  "initializer_range": 0.02,
  "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"
  ],
  "max_position_embeddings": 32768,
  "max_seq_len": 8192,
  "max_window_layers": 28,
  "model_type": "embformer",
  "num_attention_heads": 8,
  "num_hidden_layers": 16,
  "num_key_value_heads": 2,
  "rms_norm_eps": 1e-06,
  "rope_scaling": null,
  "rope_theta": 10000.0,
  "sliding_window": null,
  "tie_word_embeddings": true,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.53.0.dev0",
  "use_cache": true,
  "use_channel_shift": true,
  "use_sliding_window": false,
  "vocab_size": 6400
}