| from typing import Dict, List, Any | |
| from transformers import AutoModelForCausalLM, AutoTokenizer | |
| from peft import PeftModel | |
| import torch | |
| class EndpointHandler(): | |
| def __init__(self, path=""): | |
| model = AutoModelForCausalLM.from_pretrained("microsoft/Phi-3-mini-4k-instruct") | |
| model = PeftModel.from_pretrained(model, "srmorfi/phi3-mini-med-adapter") | |
| tokenizer = AutoTokenizer.from_pretrained("microsoft/Phi-3-mini-4k-instruct") | |
| self.model = model | |
| self.tokenizer = tokenizer | |
| self.model.eval() | |
| def __call__(self, data: Dict[str, Any]) -> List[Dict[str, Any]]: | |
| """ | |
| Process input data and generate predictions using the model. | |
| Args: | |
| data (Dict[str, Any]): Input data containing either an "inputs" key | |
| or the input directly in the data dictionary. | |
| Returns: | |
| List[Dict[str, Any]]: Processed model predictions that will be serialized and returned. | |
| """ | |
| inputs = data.pop("inputs", data) | |
| inputs = self.tokenizer(inputs, return_tensors="pt") | |
| print(inputs) | |
| with torch.no_grad(): | |
| outputs = self.model.generate(input_ids=inputs["input_ids"], max_new_tokens=20) | |
| output = self.tokenizer.batch_decode(outputs.detach().cpu().numpy(), skip_special_tokens=True)[0] | |
| predictions = [{"generated_text": output}] | |
| return predictions |