Model name update

config.py

@@ -4,6 +4,8 @@ class ConfigConstants:
     DATA_SET_NAMES = ['covidqa', 'cuad', 'delucionqa', 'emanual', 'expertqa', 'finqa', 'hagrid', 'hotpotqa', 'msmarco', 'pubmedqa', 'tatqa', 'techqa']
     EMBEDDING_MODEL_NAME = "sentence-transformers/paraphrase-MiniLM-L3-v2"
     RE_RANKER_MODEL_NAME = 'cross-encoder/ms-marco-electra-base'
+    GENERATION_MODEL_NAME = 'mixtral-8x7b-32768'
+    VALIDATION_MODEL_NAME = 'llama3-70b-8192'
     DEFAULT_CHUNK_SIZE = 1000
     CHUNK_OVERLAP = 200
 

generator/initialize_llm.py

@@ -2,9 +2,11 @@ import logging
 import os
 from langchain_groq import ChatGroq
 
+from config import ConfigConstants
+
 def initialize_generation_llm():
     os.environ["GROQ_API_KEY"] = "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"
-    model_name = "mixtral-8x7b-32768"
+    model_name = ConfigConstants.GENERATION_MODEL_NAME
     llm = ChatGroq(model=model_name, temperature=0.7)
     llm.name = model_name
     logging.info(f'Generation LLM {model_name} initialized')
@@ -12,7 +14,7 @@ def initialize_generation_llm():
 
 def initialize_validation_llm():
     os.environ["GROQ_API_KEY"] = "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"
-    model_name = "llama3-70b-8192"
+    model_name = ConfigConstants.VALIDATION_MODEL_NAME
     llm = ChatGroq(model=model_name, temperature=0.7)
     llm.name = model_name
     logging.info(f'Validation LLM {model_name} initialized')