56 lines
1.8 KiB
Python
56 lines
1.8 KiB
Python
from transformers import PretrainedConfig
|
|
|
|
|
|
class Starcoder2Config(PretrainedConfig):
|
|
model_type = "starcoder2"
|
|
keys_to_ignore_at_inference = ["past_key_values"]
|
|
|
|
def __init__(
|
|
self,
|
|
vocab_size=49152,
|
|
hidden_size=3072,
|
|
intermediate_size=12288,
|
|
num_hidden_layers=30,
|
|
num_attention_heads=24,
|
|
num_key_value_heads=2,
|
|
hidden_act="gelu_pytorch_tanh",
|
|
max_position_embeddings=4096,
|
|
initializer_range=0.018042,
|
|
norm_epsilon=1e-5,
|
|
use_cache=True,
|
|
bos_token_id=50256,
|
|
eos_token_id=50256,
|
|
rope_theta=10000.0,
|
|
sliding_window=None,
|
|
attention_dropout=0.0,
|
|
residual_dropout=0.0,
|
|
embedding_dropout=0.0,
|
|
use_bias=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.sliding_window = sliding_window
|
|
self.use_bias = use_bias
|
|
self.num_key_value_heads = num_key_value_heads
|
|
self.hidden_act = hidden_act
|
|
self.initializer_range = initializer_range
|
|
self.norm_epsilon = norm_epsilon
|
|
self.use_cache = use_cache
|
|
self.rope_theta = rope_theta
|
|
self.attention_dropout = attention_dropout
|
|
self.residual_dropout = residual_dropout
|
|
self.embedding_dropout = embedding_dropout
|
|
|
|
super().__init__(
|
|
bos_token_id=bos_token_id,
|
|
eos_token_id=eos_token_id,
|
|
**kwargs,
|
|
)
|
|
if self.architectures is None:
|
|
self.architectures = ['Starcoder2ForCausalLM']
|