feat(chatbot): Implementamos Mistal AI

Pipfile

@@ -4,8 +4,11 @@ verify_ssl = true
 name = "pypi"
 
 [packages]
-huggingface-hub = "==0.22.2"
+huggingface-hub = "*"
 gradio = "*"
+langchain-huggingface = "*"
+langchain = "*"
+langchain-core = "*"
 
 [dev-packages]
 

app.py

@@ -1,61 +1,60 @@
 import gradio as gr
-from huggingface_hub import InferenceClient
+from langchain_huggingface import HuggingFaceEndpoint
+from langchain_core.prompts import PromptTemplate
+#from langchain.globals import set_verbose, set_debug
 
+#set_verbose(True)
+#set_debug(True)
+
+repo_id = "mistralai/Mixtral-8x7B-Instruct-v0.1"
+#repo_id = "meta-llama/Meta-Llama-3.1-8B-Instruct"
+#repo_id = "HuggingFaceH4/zephyr-7b-beta"
+
+template = """[INST]Tu eres Harry Potter, el mago más hábil de todo el mundo mágico.
+Responde amablemente a la consulta del usuario basado en la información disponible.
+Si no sabes la respuesta, pide al usuario que intente reformular su consulta.
+
+{question}
+[/INST]
 """
-For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
-"""
-client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
 
+prompt = PromptTemplate.from_template(template)
+
+llm = HuggingFaceEndpoint(
+    repo_id = repo_id,
+    task = "text-generation",
+    temperature = 0.5,
+    model_kwargs = {
+        "min_length": 200,
+        "max_length": 2000,
+        "num_return_sequences": 1
+    }
+)
 
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-):
-    messages = [{"role": "system", "content": system_message}]
+llm_chain = (
+    prompt
+    | llm
+)
 
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
 
-    messages.append({"role": "user", "content": message})
+def respond(message, history):
+#    history_langchain_format = []
+#
+#    for human, ai in history:
+#        history_langchain_format.append(HumanMessage(content=human))
+#        history_langchain_format.append(AIMessage(content=ai))
+#
+#    history_langchain_format.append(HumanMessage(content=message))
 
-    response = ""
+    #print(message)
+    response = llm_chain.invoke(message)
+    #print(response)
 
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
-        temperature=temperature,
-        top_p=top_p,
-    ):
-        token = message.choices[0].delta.content
+    return response
 
-        response += token
-        yield response
 
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
 demo = gr.ChatInterface(
-    respond,
-    additional_inputs=[
-        gr.Textbox(value="Tu eres Harry Potter, el mago más hábil de todo el mundo mágico. Responde amablemente a la consulta del usuario basado en la información disponible. Si no sabes la respuesta, pide al usuario que intente reformular su consulta.", label="System message"),
-        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
-        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
-        gr.Slider(
-            minimum=0.1,
-            maximum=1.0,
-            value=0.95,
-            step=0.05,
-            label="Top-p (nucleus sampling)",
-        ),
-    ],
+    respond
 )
 
 

requirements.txt

@@ -1 +1,4 @@
-huggingface_hub==0.22.2
+huggingface_hub
+langchain-huggingface
+langchain
+langchain-core