Downloads · 30 days
10
10% of all-time downloads
opus-mt-en-tlh
MihaiPopa-1/opus-mt-en-tlh
opus-mt-en-tlh is a translation model from MihaiPopa-1. Use it when you need text moved from one language to another. The card lists the license as mit.
All-time downloads
97
Public
Parameters
82.7M
1.3 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.bin331 MB · 50%
At a glance
- Task
- Translation
- License
- mit
- Model type
- marian
- Access
- Public
- Created
- Jan 13, 2026
- Updated
- Jan 29, 2026
- SHA
- c45303d3
Try a prompt
Base models
Datasets
Model card
From the Hugging Face model README
All of this Readme is generated by Gemini 3 Flash
⚔️ Klingon-English-82.7M-Warrior
This is a lightweight, high-efficiency sequence-to-sequence model specialized for English to Klingon translation. At only 82.7M parameters, it is designed to run on extremely low-resource hardware (down to 4GB RAM) while maintaining high structural accuracy.
📊 Training Results
The model was trained for 5 epochs, achieving a highly optimized balance between learning and generalization.
| Epoch | Training Loss | Validation Loss | Status |
|---|---|---|---|
| 5 | 1.419400 | 1.239786 | 🏆 Optimized |
Note on Performance: The validation loss (1.23) is significantly lower than the training loss (1.41). This indicates the model has generalized exceptionally well and is not overfitting to the training noise.
🛠️ Technical Details
- Model Size: 82.7 Million Parameters
- Architecture: Transformer-based Encoder-Decoder
- Input: English (en)
- Output: Klingon (tlh)
- Target Hardware: CPU-friendly / Mobile / Low-RAM (4GB+)
🚀 Usage
You can use this model directly with the Hugging Face pipeline API:
from transformers import pipeline
translator = pipeline("translation", model="MihaiPopa-1/opus-mt-en-tlh")
result = translator("Glory to you and your house!")
print(result[0]['translation_text'])
- Task
- Translation
- Type
- marian
- License
- mit
- Languages
- en, tlh
- Created
- Jan 13, 2026
- Updated
- Jan 29, 2026