Create handler.py

handler.py

@@ -0,0 +1,22 @@
+from typing import Dict, List, Any
+from transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig
+import torch
+
+class EndpointHandler():
+    def __init__(self, path=""):
+      self.base_model = path
+
+      bitsandbytes= BitsAndBytesConfig(load_in_4bit=True, bnb_4bit_use_double_quant=True, bnb_4bit_quant_type="nf4", bnb_4bit_compute_dtype=torch.float16)
+      self.model = AutoModelForCausalLM.from_pretrained(self.base_model, device_map={"":0},quantization_config= bitsandbytes, trust_remote_code= True)
+      self.tokenizer = AutoTokenizer.from_pretrained(self.base_model, trust_remote_code=True)
+      self.tokenizer.pad_token = self.tokenizer.eos_token
+
+    def __call__(self, data: Dict[str, Any]) -> List[Dict[str, Any]]:
+      inputs = data.pop("inputs",data)
+      prompt = f"Below is an instruction that describes a task. Write a response that appropriately completes the request. ### Instruction: {inputs} ### Response:"
+      model_inputs = self.tokenizer([prompt], return_tensors="pt", padding=True).to("cuda")
+      generated_ids = self.model.generate(**model_inputs, max_length=200)
+      output = self.tokenizer.batch_decode(generated_ids, skip_special_tokens=True)
+      answer_without_prompt = output[0].split("### Response:")[1].strip()
+      prediction = answer_without_prompt.split("###")[0].strip()
+      return [{"generated_text": prediction}]