Incorporate high level llama api

examples/llamacpp_processor.py

@@ -1,11 +1,9 @@
 from enum import Enum
 
-import numpy as np
-from llama_cpp import Llama, LogitsProcessorList, StoppingCriteria, StoppingCriteriaList
-from numpy.typing import NDArray
+from llama_cpp import Llama, LogitsProcessorList
 from pydantic import BaseModel, constr
 
-from outlines.generate.processors.llamacpp import JSONLogitsProcessor
+from outlines.models.llamacpp import JSONLogitsProcessor
 
 
 class Weapon(str, Enum):
@@ -31,38 +29,20 @@ class Character(BaseModel):
     strength: int
 
 
-# TODO: why do we need this?
-class EosCriteria(StoppingCriteria):
-    def __init__(self, eos_token_id):
-        self.eos_token_id = eos_token_id
-
-    def __call__(self, input_ids: NDArray[np.intc], logits: NDArray[np.single]):
-        if self.eos_token_id in input_ids[1:]:
-            return True
-
-
 if __name__ == "__main__":
     llama = Llama("./phi-2.Q4_K_M.gguf")
 
-    prompt = b"Instruct: You are a leading role play gamer. You have seen thousands of different characters and their attributes.\nPlease return a JSON object with common attributes of an RPG character. Give me a character description\nOutput:"
+    prompt = "Instruct: You are a leading role play gamer. You have seen thousands of different characters and their attributes.\nPlease return a JSON object with common attributes of an RPG character. Give me a character description\nOutput:"
 
     logits_processor = JSONLogitsProcessor(Character, llama)
-    stopping_criteria_list = StoppingCriteriaList([EosCriteria(llama.token_eos())])
 
-    json_str = ""
-    tokens = llama.tokenize(prompt)
-    for token in llama.generate(
-        tokens,
+    json_str = llama.create_completion(
+        prompt,
         top_k=40,
         top_p=0.95,
-        temp=0.7,
+        temperature=0.7,
+        max_tokens=120,
         logits_processor=LogitsProcessorList([logits_processor]),
-        stopping_criteria=stopping_criteria_list,
-    ):
-        d = llama.detokenize([token])
-        try:
-            json_str += d.decode("utf-8")
-        except UnicodeDecodeError:
-            continue
+    )["choices"][0]["text"]
 
     print(json_str)