config.json

{
  "architectures": [
    "MapleForCausalLM"
  ],
  "attention_dropout": 0.0,
  "auto_map": {
    "AutoConfig": "configuration_maple.MapleConfig",
    "AutoModel": "modeling_maple.MapleModel",
    "AutoModelForCausalLM": "modeling_maple.MapleForCausalLM"
  },
  "bos_token_id": 151643,
  "dtype": "bfloat16",
  "embedding_dropout": 0.0,
  "eos_token_id": 151645,
  "head_dim": 128,
  "hidden_act": "silu",
  "hidden_size": 2048,
  "initializer_range": 0.02,
  "intermediate_size": 4096,
  "layer_types": [
    "sliding_attention",
    "sliding_attention",
    "sliding_attention",
    "full_attention",
    "sliding_attention",
    "sliding_attention",
    "sliding_attention",
    "full_attention",
    "sliding_attention",
    "sliding_attention",
    "sliding_attention",
    "full_attention",
    "sliding_attention",
    "sliding_attention",
    "sliding_attention",
    "full_attention",
    "sliding_attention",
    "sliding_attention",
    "sliding_attention",
    "full_attention",
    "sliding_attention",
    "sliding_attention",
    "sliding_attention",
    "full_attention"
  ],
  "max_position_embeddings": 131072,
  "max_window_layers": 24,
  "moe_intermediate_size": 512,
  "moe_router_enable_expert_bias": false,
  "nope_on_global_attention": true,
  "norm_topk_prob": true,
  "num_attention_heads": 16,
  "num_experts": 256,
  "num_experts_per_tok": 8,
  "num_hidden_layers": 24,
  "num_key_value_heads": 4,
  "num_shared_experts": 0,
  "output_dropout": 0.0,
  "output_router_logits": false,
  "pad_token_id": null,
  "partial_rotary_factor": 0.5,
  "preaffine": false,
  "quantize": true,
  "rms_norm_eps": 1e-06,
  "rope_scaling": null,
  "rope_theta": 10000,
  "router_dtype": "fp32",
  "sliding_window": 512,
  "tie_word_embeddings": false,
  "transformers_version": "4.57.1",
  "use_cache": true,
  "use_qk_norm": true,
  "use_rmsnorm": true,
  "vocab_size": 151936
}