Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,45 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language:
|
| 3 |
+
- en
|
| 4 |
+
tags:
|
| 5 |
+
- topic-modeling
|
| 6 |
+
- story-generation
|
| 7 |
+
- nlp
|
| 8 |
+
- pytorch
|
| 9 |
+
license: mit
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
|
| 13 |
+
# TopicNet-TopNet Integrated Model
|
| 14 |
+
|
| 15 |
+
Unified Framework for Semantic Graph-Guided Topic Discovery and Neural Topic-Driven Story Generation
|
| 16 |
+
|
| 17 |
+
## Model Architecture
|
| 18 |
+
- **TopicNet**: Semantic graph-guided topic discovery with LSTM encoder
|
| 19 |
+
- **TopNet**: Neural topic-driven story generation with cross-modal attention
|
| 20 |
+
- **Cross-Modal Attention**: Aligns topic discovery with story generation
|
| 21 |
+
|
| 22 |
+
## Training Results
|
| 23 |
+
- Topic Loss: ~2.996
|
| 24 |
+
- Story Loss: ~9.211
|
| 25 |
+
- Trained on NVIDIA V100 GPU
|
| 26 |
+
- 20 epochs, 200 batches/epoch
|
| 27 |
+
|
| 28 |
+
## Usage
|
| 29 |
+
```python
|
| 30 |
+
from transformers import AutoModel, AutoTokenizer
|
| 31 |
+
|
| 32 |
+
model = AutoModel.from_pretrained("your-username/topicnet-topnet-integrated")
|
| 33 |
+
tokenizer = AutoTokenizer.from_pretrained("your-username/topicnet-topnet-integrated")
|
| 34 |
+
```
|
| 35 |
+
|
| 36 |
+
|
| 37 |
+
## Citation
|
| 38 |
+
```bibtex
|
| 39 |
+
@article{topicnet_topnet_2024,
|
| 40 |
+
title={Unified Framework for Semantic Graph-Guided Topic Discovery and Neural Topic-Driven Story Generation},
|
| 41 |
+
author={Your Name},
|
| 42 |
+
journal={arXiv preprint},
|
| 43 |
+
year={2024}
|
| 44 |
+
}
|
| 45 |
+
```
|