File size: 1,809 Bytes
b656d19
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56

# Model Loader Script - Compatible with Python 3.13.9
import torch
import torch.nn as nn
from transformers import AutoModel, AutoTokenizer
import json
import os

class MultiHeadClassifier(nn.Module):
    def __init__(self, bert_model_name="aubmindlab/bert-base-arabertv2", num_heads=5, num_classes=4):
        super().__init__()
        self.bert = AutoModel.from_pretrained(bert_model_name)
        hidden_size = self.bert.config.hidden_size
        self.heads = nn.ModuleList([
            nn.Sequential(
                nn.Linear(hidden_size, 256),
                nn.ReLU(),
                nn.Dropout(0.3),
                nn.Linear(256, num_classes)
            ) for _ in range(num_heads)
        ])

    def forward(self, input_ids, attention_mask):
        outputs = self.bert(input_ids=input_ids, attention_mask=attention_mask)
        cls_embedding = outputs.last_hidden_state[:, 0, :]
        logits = [head(cls_embedding) for head in self.heads]
        return logits

def load_feedback_model(model_path, tokenizer_path):
    """Load the trained model for inference"""
    # Load config
    with open(os.path.join(os.path.dirname(model_path), "model_config.json"), "r") as f:
        config = json.load(f)

    # Initialize model
    model = MultiHeadClassifier(
        bert_model_name=config["bert_model_name"],
        num_heads=config["num_heads"],
        num_classes=config["num_classes"]
    )

    # Load weights
    model.load_state_dict(torch.load(model_path, map_location="cpu"))
    model.eval()

    # Load tokenizer
    tokenizer = AutoTokenizer.from_pretrained(tokenizer_path)

    return model, tokenizer, config

# Example usage:
# model, tokenizer, config = load_feedback_model(
#     "saved_model_python313/feedback_classifier.pt",
#     "saved_model_python313/tokenizer"
# )