Ziya-LLaMA-13B-v1.1/bert4torch_config.json

{
"model": "llama",
"bos_token_id": 1,
"eos_token_id": 2,
"hidden_act": "silu",
"hidden_size": 5120,
"initializer_range": 0.02,
"intermediate_size": 13824,
"max_position_embeddings": 2048,
"num_attention_heads": 40,
"num_hidden_layers": 40,
"pad_token_id": 0,
"layer_norm_eps": 1e-06,
"tie_word_embeddings": false,
"use_cache": true,
"vocab_size": 39424,
"segment_vocab_size": 0,
"skip_init": true,
"rope_rank": "updown"
}