As a user, I want a method for translating an input text using the NLLB-200 model, so that I can generate translations for evaluation.

[laurejt]

Needs to support Chinese, Japanese, English, and Spanish but as possible make it easy to support other model-supported languages.

Use the NLLB-200-3.3B HuggingFace model.

Inputs:

• src_lang: str: Source language. Use ISO 639 language codes (e.g., "zh", "ja", "es", "en")
• tgt_lang: str: Target language. Use ISO 639 language codes
• text: str: Text to be translated from the source to target language

Output: The translated text as a string.

Note: For this model you will need to make a mapping from the input language codes to the codes used by NLLB-200 (see Table 1 in the research paper). For now, assume Chinese should map to simplified Chinese.