metadata

license: apache-2.0
language:
  - en
  - ja
base_model:
  - utter-project/EuroLLM-9B

Description

This is a sentence-based English-to-Japanese translation model fine-tuned from utter-project/EuroLLM-9B.

Training data: WMT2025 en-ja parallel corpus

Example

import transformers
import torch

model_id = "Systran/EuroLLM-9B-cpt-ft-wmt25-en-ja"

pipeline = transformers.pipeline(
    "text-generation",
    model=model_id,
    model_kwargs={"torch_dtype": torch.bfloat16},
    device_map="auto",
)

messages = "Please translate the following English text into Japanese.\nInput: Hello World!\nOutput: "

outputs = pipeline(
    messages,
    max_new_tokens=256,
)
print(outputs[0]["generated_text"][len(messages):].split("\n")[0])