mirror of
https://github.com/datawhalechina/llms-from-scratch-cn.git
synced 2026-10-11 03:27:25 +00:00
76 lines
2.4 KiB
Python
76 lines
2.4 KiB
Python
import logging
|
|
import json
|
|
|
|
logger = logging.getLogger(__name__)
|
|
|
|
class MiniCPMConfig():
|
|
|
|
model_type = "minicpm"
|
|
keys_to_ignore_at_inference = ["past_key_values"]
|
|
|
|
def __init__(
|
|
self,
|
|
vocab_size=32000,
|
|
hidden_size=4096,
|
|
intermediate_size=11008,
|
|
num_hidden_layers=32,
|
|
num_attention_heads=32,
|
|
num_key_value_heads=None,
|
|
hidden_act="silu",
|
|
max_position_embeddings=2048,
|
|
initializer_range=0.02,
|
|
rms_norm_eps=1e-6,
|
|
use_cache=False,
|
|
pad_token_id=None,
|
|
bos_token_id=1,
|
|
eos_token_id=2,
|
|
pretraining_tp=1,
|
|
tie_word_embeddings=True,
|
|
rope_theta=10000.0,
|
|
rope_scaling=None,
|
|
attention_bias=False,
|
|
attention_dropout=0.0,
|
|
scale_emb=1,
|
|
dim_model_base=1,
|
|
scale_depth=1,
|
|
output_attentions=False,
|
|
output_hidden_states=False,
|
|
return_dict=True,
|
|
use_return_dict=True,
|
|
**kwargs,
|
|
):
|
|
self.vocab_size = vocab_size
|
|
self.max_position_embeddings = max_position_embeddings
|
|
self.hidden_size = hidden_size
|
|
self.intermediate_size = intermediate_size
|
|
self.num_hidden_layers = num_hidden_layers
|
|
self.num_attention_heads = num_attention_heads
|
|
self.num_key_value_heads = num_key_value_heads
|
|
self.hidden_act = hidden_act
|
|
self.initializer_range = initializer_range
|
|
self.rms_norm_eps = rms_norm_eps
|
|
self.pretraining_tp = pretraining_tp
|
|
self.use_cache = use_cache
|
|
self.rope_theta = rope_theta
|
|
self.rope_scaling = rope_scaling
|
|
self.attention_bias = attention_bias
|
|
self.attention_dropout = attention_dropout
|
|
self.scale_emb = scale_emb
|
|
self.dim_model_base = dim_model_base
|
|
self.scale_depth = scale_depth
|
|
self.pad_token_id=pad_token_id
|
|
self.bos_token_id=bos_token_id
|
|
self.eos_token_id=eos_token_id
|
|
self.tie_word_embeddings=tie_word_embeddings
|
|
self.output_attentions=output_attentions
|
|
self.output_hidden_states=output_hidden_states
|
|
self.return_dict=return_dict
|
|
self.use_return_dict=use_return_dict
|
|
|
|
def to_json_string(self) -> str:
|
|
config_dict = self.__dict__
|
|
return json.dumps(config_dict, indent=2, sort_keys=True) + "\n"
|
|
|
|
def __repr__(self):
|
|
return f"{self.__class__.__name__} {self.to_json_string()}"
|