Karez's picture
Update Baseline-Word-Model/README.md
c0c5094 verified
|
Raw
History Blame Contribute Delete
771 Bytes
# Baseline CRNN (No Attention)
CRNN baseline without any attention block.
## Architecture
- Family: **Baseline**
- CNN backbone: 6 convolutional blocks (max 256 channels)
- Recurrent block: 3 BiLSTM layers, hidden = 160 per direction
- Decoding: Connectionist Temporal Classification (CTC)
- Parameters: 3,838,065
## Test Performance (DASTNUS, seed 42)
| Metric | Value |
|--------|-------|
| Test CER | 0.0380 |
| Test WER | 0.1544 |
## Files
- `model.safetensors` — model weights
- `config.json` — architecture and training configuration
- `vocab.json` — character-to-index mapping (CTC blank at index 0)
- `idx_to_char.json` — reverse mapping for decoding
## License
Released for non-commercial scientific research purposes only.