Method load

cmd/modeldeployer/main.py:21–25

(self)

        self.ready = False

    def load(self):
        self.tokenizer = AutoTokenizer.from_pretrained(f"{self.org_name}/{self.repo_name}", trust_remote_code=True)
        self.model = SentenceTransformer(f"{self.org_name}/{self.repo_name}", trust_remote_code=True)
        self.model.max_seq_length = self.max_sequence_length
        self.ready = True

    def predict(self, payload: Dict, headers: Dict) -> Dict:
        inputs = payload["instances"]