Migramos a Groq

Pipfile

@@ -12,6 +12,7 @@ langchain-core = "*"
 transformers = "*"
 llama-cpp-python = "*"
 langchain-community = "*"
+langchain-groq = "*"
 
 [dev-packages]
 

app.py

@@ -1,6 +1,5 @@
 import gradio as gr
-from huggingface_hub import hf_hub_download
-from langchain_community.chat_models import ChatLlamaCpp
+from langchain_groq import ChatGroq
 from langchain_core.prompts import ChatPromptTemplate
 from langchain.globals import set_verbose, set_debug
 import os
@@ -31,20 +30,17 @@ def initLLM():
     """
     Inicializamos el modelo LLM.
 
-    Otros modelos que podríamos usar:
-        * bartowski/Llama-3.2-1B-Instruct-GGUF
-        * HuggingFaceH4/zephyr-7b-beta
+    Modelos:
+        * llama3-8b-8192
+        * mixtral-8x7b-32768
     """
-
-    model_path = hf_hub_download(repo_id="bartowski/Llama-3.2-1B-Instruct-GGUF", filename="Llama-3.2-1B-Instruct-IQ4_XS.gguf")
-
-    llm = ChatLlamaCpp(
-        model_path=model_path,
-        temperature=0.7,
-        max_tokens=500,
-        top_p=1,
-        # callback_manager=callback_manager,
-        # verbose=True,  # Verbose is required to pass to the callback manager
+    llm = ChatGroq(
+        model="mixtral-8x7b-32768",
+        temperature=0,
+        max_tokens=None,
+        timeout=None,
+        max_retries=2,
+        # other params...
     )
 
     return llm

requirements.txt

@@ -1,6 +1,3 @@
-huggingface-hub
-langchain-huggingface
 langchain
 langchain-core
-langchain-community
-llama-cpp-python
+langchain-groq