This model Confused404/eng-gmq-finetuned_v2-no is a fine-tuned version of the Helsinki-NLP/opus-mt-en-gmq model on a combination of the following datasets:

CCMatrix (English-Norwegian parallel sentences)
OpenSubtitles (informal and conversational language pairs)

The goal of this model is to improve English → Norwegian Bokmål (ISO 639-1: no) translation accuracy for both formal and informal domains, surpassing the baseline Helsinki-NLP general model.

🛠️ How to Use

python

from transformers import MarianMTModel, MarianTokenizer

model_name = "Confused404/eng-gmq-finetuned_v2-no"
tokenizer = MarianTokenizer.from_pretrained(model_name)
model = MarianMTModel.from_pretrained(model_name)

def translate(texts):
    inputs = tokenizer(texts, return_tensors="pt", padding=True, truncation=True)
    translated = model.generate(**inputs)
    return [tokenizer.decode(t, skip_special_tokens=True) for t in translated]

translate(["Hello, how are you?", "This is a test sentence."])