Meta-Llama-3-8B/bert4torch_config.json

{
	"model": "llama",
	"hidden_size": 4096,
    "intermediate_size": 14336, 
	"num_attention_heads": 32,
	"num_hidden_layers": 32,
	"hidden_act": "silu",
	"vocab_size": 128256,
	"segment_vocab_size": 0,
	"skip_init": true,
	"layer_norm_eps": 1e-05,
	"rope_rank": "updown",
	"rope_theta": 500000.0,
	"num_key_value_heads": 8,
	"max_position_embeddings": 8192,
	"torch_dtype": "bfloat16",
	"tie_word_embeddings": false,
	"attention_probs_dropout_prob": 0.0
}