StrongTextCNN
A CNN-based text classification model trained using PyTorch.
Model Architecture
- Embedding Layer (300 dimensions)
- Conv1D kernels: 2, 3, 4, 5
- 128 filters each
- Batch Normalization
- Fully Connected Layers:
- 512 → 256
- 256 → 64
- 64 → 1
- Dropout = 0.4
Files
- model.py
- best_model.pt
- tokenizer/
Usage
from model import StrongTextCNN