33 lines
825 B
JSON
33 lines
825 B
JSON
|
{
|
||
|
|
"_name_or_path": "sapienzanlp/modello-italia-9b",
|
||
|
|
"architectures": [
|
||
|
|
"GPTNeoXForCausalLM"
|
||
|
|
],
|
||
|
|
"attention_bias": true,
|
||
|
|
"attention_dropout": 0.0,
|
||
|
|
"attention_probs_dropout_prob": 0,
|
||
|
|
"bos_token_id": 0,
|
||
|
|
"classifier_dropout": 0.1,
|
||
|
|
"eos_token_id": 0,
|
||
|
|
"hidden_act": "gelu_fast",
|
||
|
|
"hidden_dropout": 0.0,
|
||
|
|
"hidden_dropout_prob": 0,
|
||
|
|
"hidden_size": 5120,
|
||
|
|
"initializer_range": 0.02,
|
||
|
|
"intermediate_size": 12800,
|
||
|
|
"layer_norm_eps": 1e-05,
|
||
|
|
"max_position_embeddings": 4096,
|
||
|
|
"model_type": "gpt_neox",
|
||
|
|
"num_attention_heads": 32,
|
||
|
|
"num_hidden_layers": 34,
|
||
|
|
"rope_scaling": null,
|
||
|
|
"rotary_emb_base": 10000,
|
||
|
|
"rotary_pct": 0.4,
|
||
|
|
"tie_word_embeddings": false,
|
||
|
|
"torch_dtype": "bfloat16",
|
||
|
|
"transformers_version": "4.40.1",
|
||
|
|
"use_cache": true,
|
||
|
|
"use_parallel_residual": true,
|
||
|
|
"vocab_size": 50176
|
||
|
|
}
|