tinystories_zh / config.json
Qi Wang
Upload 3 files
87e5187
raw
history blame
618 Bytes
{
"_name_or_path": "../models/tinystories",
"architectures": [
"LlamaForCausalLM"
],
"bos_token_id": 1,
"eos_token_id": 2,
"hidden_act": "silu",
"hidden_size": 768,
"initializer_range": 0.02,
"intermediate_size": 2268,
"max_position_embeddings": 768,
"model_type": "llama",
"num_attention_heads": 12,
"num_hidden_layers": 12,
"num_key_value_heads": 12,
"pad_token_id": 3,
"pretraining_tp": 1,
"rms_norm_eps": 1e-06,
"rope_scaling": null,
"tie_word_embeddings": false,
"torch_dtype": "float16",
"transformers_version": "4.32.1",
"use_cache": true,
"vocab_size": 3000
}