{
  "architectures": [
    "BuddyGPTForCausalLM"
  ],
  "attention_dropout": 0.0,
  "attn_impl": "mla",
  "eos_token_id": 151645,
  "hidden_act": "silu",
  "hidden_size": 1536,
  "initializer_range": 0.02,
  "intermediate_size": 3072,
  "kv_lora_rank": 16,
  "model_type": "buddygpt",
  "moe_intermediate_size": 256,
  "n_expert": 12,
  "n_expert_per_token": 2,
  "n_group": 2,
  "n_shared_experts": 2,
  "n_topk_group": 1,
  "norm_topk_prob": true,
  "num_attention_heads": 16,
  "num_hidden_layers": 24,
  "num_key_value_heads": 8,
  "num_seq_len": 1024,
  "pad_token_id": 151643,
  "q_lora_rank": 16,
  "qk_nope_head_dim": 72,
  "qk_rope_head_dim": 24,
  "rms_norm_eps": 1e-06,
  "rope_theta": 100000.0,
  "routed_scaling_factor": 0.2,
  "scoring_func": "sigmoid",
  "tie_word_embeddings": true,
  "topk_method": "noaux_tc",
  "torch_dtype": "float32",
  "transformers_version": "4.52.4",
  "use_cache": true,
  "v_head_dim": 96,
  "vocab_size": 151669
}