deploy 3

api/question_answering/qa_model.py

@@ -13,12 +13,11 @@ from langchain.llms import HuggingFacePipeline
 from langchain.llms.base import LLM
 from langchain.embeddings import HuggingFaceEmbeddings, HuggingFaceHubEmbeddings, HuggingFaceInstructEmbeddings
 from langchain.vectorstores import FAISS
-from llama_cpp import Llama
 
 from api.logger import logger
 from api.question_answering.response import Response
 
-
+"""
 class LocalBinaryModel(LLM):
     model_id: str = None
     llm: Llama = None
@@ -49,7 +48,7 @@ class LocalBinaryModel(LLM):
     @property
     def _llm_type(self) -> str:
         return self.model_id
-
+"""
 
 class TransformersPipelineModel(LLM):
     model_id: str = None
@@ -170,9 +169,9 @@ class QAModel():
 
         if 'local_models/' in llm_model_id:
             logger.info('using local binary model')
-            self.llm_model = LocalBinaryModel(
-                model_id=llm_model_id
-            )
+            #self.llm_model = LocalBinaryModel(
+            #    model_id=llm_model_id
+            #)
         elif 'api_models/' in llm_model_id:
             logger.info('using api served model')
             self.llm_model = APIServedModel(

requirements.txt

@@ -7,7 +7,6 @@ discord.py==2.2.2
 evaluate==0.4.0
 fastapi==0.98.0
 langchain==0.0.154
-llama_cpp_python==0.1.39
 nltk==3.8.1
 nbconvert==7.6.0
 nbformat==5.9.0