Switch from InferenceClient to local model loading with SmolLM2-1.7B-Instruct

- Replace HuggingFace Inference API with local model loading
- Use transformers library to load SmolLM2-1.7B-Instruct model
- Update prompt formatting for local model generation
- Modify response generation to work with local model
- Update requirements.txt with necessary dependencies

app.py

@@ -1,11 +1,19 @@
 import gradio as gr
-from huggingface_hub import InferenceClient
+from transformers import AutoModelForCausalLM, AutoTokenizer
+import torch
 
 """
 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")
 
+# Cargar el modelo y el tokenizer
+model_name = "HuggingFaceTB/SmolLM2-1.7B-Instruct"
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+model = AutoModelForCausalLM.from_pretrained(
+    model_name,
+    torch_dtype=torch.float16,
+    device_map="auto"
+)
 
 def respond(
     message,
@@ -15,29 +23,37 @@ def respond(
     temperature,
     top_p,
 ):
-    messages = [{"role": "system", "content": system_message}]
-
+    # Construir el prompt con el formato correcto
+    prompt = f"<|system|>\n{system_message}</s>\n"
+    
     for val in history:
         if val[0]:
-            messages.append({"role": "user", "content": val[0]})
+            prompt += f"<|user|>\n{val[0]}</s>\n"
         if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
-
-    messages.append({"role": "user", "content": message})
-
-    response = ""
-
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
+            prompt += f"<|assistant|>\n{val[1]}</s>\n"
+    
+    prompt += f"<|user|>\n{message}</s>\n<|assistant|>\n"
+    
+    # Tokenizar el prompt
+    inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
+    
+    # Generar la respuesta
+    outputs = model.generate(
+        **inputs,
+        max_new_tokens=max_tokens,
         temperature=temperature,
         top_p=top_p,
-    ):
-        token = message.choices[0].delta.content
-
-        response += token
-        yield response
+        do_sample=True,
+        pad_token_id=tokenizer.eos_token_id
+    )
+    
+    # Decodificar la respuesta
+    response = tokenizer.decode(outputs[0], skip_special_tokens=True)
+    
+    # Extraer solo la parte de la respuesta del asistente
+    response = response.split("<|assistant|>\n")[-1].strip()
+    
+    yield response
 
 
 """

requirements.txt

@@ -1 +1,5 @@
-huggingface_hub==0.25.2
+huggingface_hub==0.25.2
+gradio>=4.0.0
+transformers>=4.36.0
+torch>=2.0.0
+accelerate>=0.25.0