Switch to better working llamacpp provider (#289)

provider/llamacpp.py

@@ -1,7 +1,7 @@
 try:
-    from pyllamacpp.model import Model
+    from llama_cpp import Llama
 except:
-    print("Failed to import pyllamacpp.")
+    print("Failed to import llama-cpp-python.")
 
 
 class LlamacppProvider:
@@ -13,18 +13,20 @@ def __init__(
         AI_MODEL: str = "default",
         **kwargs
     ):
-        self.requirements = ["pyllamacpp"]
+        self.requirements = ["llama-cpp-python"]
         self.AI_TEMPERATURE = AI_TEMPERATURE
         self.MAX_TOKENS = MAX_TOKENS
         self.AI_MODEL = AI_MODEL
+
         if MODEL_PATH:
             try:
                 self.MAX_TOKENS = int(self.MAX_TOKENS)
             except:
                 self.MAX_TOKENS = 2000
-            self.model = Model(ggml_model=MODEL_PATH, n_ctx=self.MAX_TOKENS)
+
+            self.model = Llama(model_path=MODEL_PATH)
 
     def instruct(self, prompt, tokens: int = 0):
-        return self.model.generate(
-            prompt, n_predict=55, n_threads=8, temp=float(self.AI_TEMPERATURE)
-        )
+        return self.model(
+            prompt, max_tokens=55, stop=["\n"], temperature=float(self.AI_TEMPERATURE)
+        )["choices"][0]["text"]

requirements-mac.txt

@@ -34,4 +34,5 @@ webdriver_manager
 spacy
 discord
 EdgeGPT
-nomic
+nomic
+llama-cpp-python==0.1.48

requirements.txt

@@ -31,7 +31,7 @@ sentence-transformers
 tiktoken
 tweepy
 webdriver_manager
-pyllamacpp==1.0.7
+llama-cpp-python==0.1.48
 spacy
 discord
 gpt4free