---
license: mit
language:
- en
- de
library_name: pytorch
pipeline_tag: translation
tags:
- pytorch
- gru
- seq2seq
- luong
- machine-translation
- multi30k
datasets:
- bentrevett/multi30k
---

# Seq2Seq GRU Translator (English -> German)

A PyTorch Seq2Seq Encoder-Decoder GRU with Luong Attention trained from scratch on the Multi30k dataset.

## Model Details

- Architecture: Seq2Seq Encoder-Decoder GRU with Luong Attention
- Framework: PyTorch
- Dataset: Multi30k (English -> German)
- Embedding Dimension: 128
- Hidden Dimension: 256

## Training

- Optimizer: Adam
- Learning Rate: 1e-3
- Loss: CrossEntropyLoss (ignore_index=PAD)
- Gradient Clipping: max_norm=1.0
- Epochs: 50
- Batch Size: 64

## Dataset

| Split | Samples |
|------|--------:|
| Train | 29,000 |
| Validation | 1,014 |
| Test | 1,000 |

## Performance

| Metric | Value |
|------|------:|
| Train Loss | 0.42 |
| Validation Loss | 5.17 |
| Test Loss | 3.26 |
| BLEU | 27.13 |

![](en_de_luong_loss.png)

## Usage

```python
# Load model
model = Seq2Seq(encoder, decoder, device)

model.load_state_dict(
    load_file("best_seq2seq_luong.safetensors")
)

model.eval()

# Inference
sentence = "A man is riding a bicycle."

translation = translate_sentence(
    sentence,
    model
)

print(" ".join(translation))
```

Author
Ankit Bari

GitHub: https://github.com/aijadugar
Hugging Face: https://huggingface.co/aijadugar