base llemma-7b and custom handler.py

handler.py

@@ -0,0 +1,37 @@
+from typing import Dict, List, Any
+from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline, StoppingCriteria, StoppingCriteriaList
+
+
+
+class EndpointHandler():
+    def __init__(self, path=""):
+        # Preload all the elements you are going to need at inference.
+        tokenizer = AutoTokenizer.from_pretrained(path)
+        model = AutoModelForCausalLM.from_pretrained(path)
+        tokenizer.pad_token = tokenizer.eos_token
+        self.pipeline = pipeline('text-generation', model=model, tokenizer=tokenizer)
+        self.stopping_criteria = StoppingCriteriaList([StopAtPeriodCriteria(tokenizer)])
+
+    def __call__(self, data: Dict[str, Any]) -> List[Dict[str, Any]]:
+        """
+       data args:
+            inputs (:obj: `str` | `PIL.Image` | `np.array`)
+            kwargs
+      Return:
+            A :obj:`list` | `dict`: will be serialized and returned
+        """
+        inputs = data.pop("inputs", data)
+
+        prediction = self.pipeline(inputs, stopping_criteria=self.stopping_criteria, max_new_tokens=100)
+        return prediction
+
+
+class StopAtPeriodCriteria(StoppingCriteria):
+    def __init__(self, tokenizer):
+        self.tokenizer = tokenizer
+
+    def __call__(self, input_ids, scores, **kwargs):
+        # Decode the last generated token to text
+        last_token_text = self.tokenizer.decode(input_ids[:, -1], skip_special_tokens=True)
+        # Check if the decoded text ends with a period
+        return '.' in last_token_text

sample.py

@@ -0,0 +1,13 @@
+from handler import EndpointHandler
+
+# init handler
+my_handler = EndpointHandler(path=".")
+
+# prepare sample payload
+payload = {"inputs": "I am so happy to reach in my pocket and find a"}
+
+# test the handler
+payload=my_handler(payload)
+
+# show results
+print("output:", payload)