File size: 1,542 Bytes
59c5c0d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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

---

language: en
tags:
- scientific-papers
- arxiv
- classification
- scibert
- text-classification
datasets:
- arxiv
license: mit
---


# SciBERT Fine-tuned for arXiv Paper Classification

This model is a fine-tuned version of [allenai/scibert_scivocab_uncased](https://huggingface.co/allenai/scibert_scivocab_uncased) for classifying scientific papers into arXiv categories.

## Model Description

- **Base Model:** SciBERT (Scientific BERT)
- **Task:** Multi-class Text Classification
- **Training Data:** arXiv scientific papers
- **Number of Classes:** 20 arXiv categories

## Intended Use

This model classifies scientific paper abstracts into their primary arXiv subject categories.

## Usage

```python

from transformers import AutoTokenizer, AutoModelForSequenceClassification

import torch



tokenizer = AutoTokenizer.from_pretrained("YOUR_USERNAME/scibert-arxiv-classifier")

model = AutoModelForSequenceClassification.from_pretrained("YOUR_USERNAME/scibert-arxiv-classifier")



text = "Your scientific paper abstract here..."

inputs = tokenizer(text, return_tensors="pt", truncation=True, max_length=512)

outputs = model(**inputs)

predictions = torch.nn.functional.softmax(outputs.logits, dim=-1)

predicted_class = torch.argmax(predictions, dim=-1).item()

```

## Training Details

- Fine-tuned on arXiv paper dataset
- Optimized for scientific domain text classification

## Limitations

- Best suited for scientific/academic papers
- Performance may vary on non-scientific text