from typing import Any, Optional from smolagents.tools import Tool class TranslationTool(Tool): """ Example: ```py translator = TranslationTool() translator("This is a super nice API!", src_lang="English", tgt_lang="French") ``` """ default_checkpoint = "facebook/nllb-200-distilled-600M" description = "This is a tool that translates text from a language to another." name = "translator" inputs = {'text': {'type': 'string', 'description': 'The text to translate'}, 'src_lang': {'type': 'string', 'description': "The language of the text to translate. Written in plain English, such as 'Romanian', or 'Albanian'"}, 'tgt_lang': {'type': 'string', 'description': "The language for the desired output language. Written in plain English, such as 'Romanian', or 'Albanian'"}} output_type = "string" def __init__(self, lang_to_code=LANGUAGE_CODES, pre_processor_class=AutoTokenizer, model_class=AutoModelForSeq2SeqLM): super().__init__() self.lang_to_code = lang_to_code self.pre_processor_class = pre_processor_class self.model_class = model_class # self.pre_processor = self.pre_processor_class.from_pretrained(self.default_checkpoint) # self.model = self.model_class.from_pretrained(self.default_checkpoint) # self.post_processor = self.pre_processor def encode(self, text, src_lang, tgt_lang): if src_lang not in self.lang_to_code: raise ValueError(f"{src_lang} is not a supported language.") if tgt_lang not in self.lang_to_code: raise ValueError(f"{tgt_lang} is not a supported language.") src_lang = self.lang_to_code[src_lang] tgt_lang = self.lang_to_code[tgt_lang] return self.pre_processor._build_translation_inputs( text, return_tensors="pt", src_lang=src_lang, tgt_lang=tgt_lang ) def forward(self, inputs): return self.model.generate(**inputs) def decode(self, outputs): return self.post_processor.decode(outputs[0].tolist(), skip_special_tokens=True)