phi-2 / config.json
susnato's picture
partial_rotary_factor : 05->0.4
f3d2d29
raw
history blame
654 Bytes
{
"architectures": [
"PhiForCausalLM"
],
"bos_token_id": 1,
"eos_token_id": 2,
"hidden_act": "gelu_new",
"hidden_size": 2560,
"initializer_range": 0.02,
"intermediate_size": 10240,
"max_position_embeddings": 2048,
"model_type": "phi",
"num_attention_heads": 32,
"num_hidden_layers": 32,
"pretraining_tp": 1,
"resid_pdrop": 0.1,
"embd_pdrop": 0.0,
"layer_norm_eps": 1e-05,
"rope_scaling": null,
"rope_theta": 10000.0,
"partial_rotary_factor": 0.4,
"qk_layernorm": false,
"tie_word_embeddings": false,
"torch_dtype": "float16",
"transformers_version": "4.35.2",
"use_cache": true,
"vocab_size": 51200
}