config.json

{
  "_name_or_path": "crumb/Llama-p-small",
  "activation_function": "silu",
  "architectures": [
    "PlusModelForCausalLM"
  ],
  "auto_map": {
    "AutoConfig": "crumb/Llama-p-small--configuration_nano.NanoConfig",
    "AutoModelForCausalLM": "crumb/Llama-p-small--modeling_nano.PlusModelForCausalLM"
  },
  "bos_token_id": 1,
  "combined_qkv": true,
  "eos_token_id": 2,
  "expanded_lm_head_size": 8192,
  "expanded_wte_size": 8192,
  "experimental_full_adaption_rank": null,
  "ffn": "llamalike",
  "full_adaptation_has_pre_proj": false,
  "full_adaptation_type": "no",
  "hidden_size": 768,
  "initializer_range": 0.02,
  "intermediate_size": 3072,
  "kv_hidden_size": null,
  "layer_norm_epsilon": 1e-06,
  "layernorm": "llamarmsnorm",
  "lm_head_bias": false,
  "lm_head_projection_bias": false,
  "max_position_embeddings": 2048,
  "model_type": "nano",
  "num_attention_heads": 12,
  "num_hidden_layers": 10,
  "pre_proj_dim": null,
  "residual_alpha": false,
  "rope_scaling": null,
  "rope_theta": 10000,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.36.2",
  "use_bias": false,
  "use_cache": false,
  "vocab_size": 32000
}