{
  "_name_or_path": "NanoChat-0.3B-base/",
  "architectures": [
    "NanoChatLM"
  ],
  "attention": "GQA",
  "auto_map": {
    "AutoConfig": "LMConfig.LMConfig",
    "AutoModel": "model.NanoChatLM",
    "AutoModelForCausalLM": "model.NanoChatLM"
  },
  "aux_loss_alpha": 0.1,
  "dim": 896,
  "dropout": 0.0,
  "flash_attn": true,
  "hidden_dim": 2432,
  "kv_lora_rank": 512,
  "max_seq_len": 512,
  "model_type": "nanochat",
  "multiple_of": 64,
  "n_heads": 14,
  "n_kv_heads": 2,
  "n_layers": 24,
  "n_routed_experts": 4,
  "n_shared_experts": true,
  "norm_eps": 1e-05,
  "norm_topk_prob": true,
  "num_experts_per_tok": 2,
  "q_lora_rank": 0,
  "qk_nope_head_dim": 64,
  "qk_rope_head_dim": 64,
  "rope_theta": 1000000.0,
  "scoring_func": "softmax",
  "seq_aux": true,
  "torch_dtype": "float32",
  "transformers_version": "4.42.4",
  "use_moe": false,
  "v_head_dim": 64,
  "vocab_size": 151650
}