| from typing import List |
| from transformers import PretrainedConfig |
|
|
| class LidirlCNNConfig(PretrainedConfig): |
| model_type = "LidirlCNN" |
| |
| def __init__(self, |
| embed_dim : int = 32, |
| channels : List[int] = [32], |
| kernels : List[int] = [3], |
| strides : List[int] = [1], |
| vocab_size: int = 256, |
| label_size : int = 200, |
| max_length : int = 1024, |
| multilabel : bool = False, |
| montecarlo_layer : bool = False, |
| **kwargs, |
| ): |
| self.embed_dim = embed_dim |
| self.channels = channels |
| self.kernels = kernels |
| self.strides = strides |
|
|
| self.vocab_size = vocab_size |
| self.label_size = label_size |
| self.max_length = max_length |
| self.multilabel = multilabel |
| self.montecarlo_layer = montecarlo_layer |
| |
| super().__init__(**kwargs) |
