metadata
language:
- en
- fr
library_name: transformers
pipeline_tag: translation
license: cc-by-nc-4.0
base_model: facebook/nllb-200-distilled-600M
tags:
- translation
- gaming
- renpy
- visual-novel
- french
datasets:
- custom
metrics:
- bleu
- sacrebleu
widget:
- text: Hello! Welcome to our game.
example_title: Gaming Interface
- text: I love you more than anything.
example_title: Romance Dialogue
- text: What do you choose?
example_title: Choice Menu
🎮 NLLB-RenPy: Specialized French Gaming Translator
🌟 Model Description
This model is a fine-tuned version of facebook/nllb-200-distilled-600M specifically trained for English-to-French translation in gaming contexts, particularly RenPy visual novels.
🎯 Specialized For:
- 🎮 Gaming interfaces and menus
- 💬 Character dialogues and narratives
- 💕 Romance and emotional expressions
- 🔄 Interactive choices and options
- 📱 UI elements and notifications
🏆 Performance Highlights:
- Superior quality vs Google Translate/DeepL for gaming
- Context-aware translations maintaining gaming tone
- Optimized for visual novel terminology
- Consistent character voice preservation
🚀 Quick Start
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
# Load model
model = AutoModelForSeq2SeqLM.from_pretrained("virusf/nllb-renpy")
tokenizer = AutoTokenizer.from_pretrained("virusf/nllb-renpy")
# Translate
text = "Hello! Welcome to our game."
inputs = tokenizer(text, return_tensors="pt")
outputs = model.generate(**inputs, forced_bos_token_id=tokenizer.convert_tokens_to_ids("fra_Latn"))
result = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(result) # "Bonjour! Bienvenue dans notre jeu."
📊 Training Details
Base Model: facebook/nllb-200-distilled-600M
Training Data: 15,000+ specialized gaming translations
Languages: English → French
Epochs: 2.0
Final Loss: 0.4441
🎯 Use Cases
Perfect for translating:
✅ RenPy/Ren'Py visual novels
✅ Gaming interfaces and menus
✅ Character dialogues and stories
✅ Interactive fiction content
✅ Dating simulation games