source/model.py

from __future__ import annotations

from transformers import BertConfig, BertForSequenceClassification


def build_model(
    vocab_size: int,
    pad_token_id: int,
) -> BertForSequenceClassification:
    config = BertConfig(
        vocab_size=vocab_size,
        hidden_size=64,
        num_hidden_layers=2,
        num_attention_heads=2,
        intermediate_size=160,
        hidden_act="gelu",
        hidden_dropout_prob=0.08,
        attention_probs_dropout_prob=0.05,
        max_position_embeddings=96,
        type_vocab_size=1,
        pad_token_id=pad_token_id,
        num_labels=2,
        id2label={0: "ROUTINE", 1: "HAZARDOUS"},
        label2id={"ROUTINE": 0, "HAZARDOUS": 1},
    )
    return BertForSequenceClassification(config)


def parameter_count(model) -> int:
    return sum(parameter.numel() for parameter in model.parameters())