File size: 856 Bytes
2fc050b |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 |
from transformers import AutoModel, AutoTokenizer
import torch.nn as nn
class MultiTaskBERT(nn.Module):
def __init__(self, num_lang_classes=4, num_sentiment_classes=2):
super().__init__()
self.bert = AutoModel.from_pretrained("distilbert-base-multilingual-cased")
self.dropout = nn.Dropout(0.3)
self.sentiment_head = nn.Linear(768, num_sentiment_classes)
self.lang_head = nn.Linear(768, num_lang_classes)
def forward(self, input_ids, attention_mask):
outputs = self.bert(input_ids=input_ids, attention_mask=attention_mask)
pooled_output = outputs.last_hidden_state[:, 0]
pooled_output = self.dropout(pooled_output)
sentiment_logits = self.sentiment_head(pooled_output)
lang_logits = self.lang_head(pooled_output)
return sentiment_logits, lang_logits |