dev

app.py

@@ -2,40 +2,41 @@ import gradio as gr
 import torch
 from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline
 
-# ✅ Recommended model for CPU systems (2vCPU / 16GB RAM)
+# Recommended for 2vCPU/16GB RAM CPU hosting
 model_id = "microsoft/phi-2"
 
-# Load tokenizer and model
+# Load model and tokenizer
 tokenizer = AutoTokenizer.from_pretrained(model_id)
 model = AutoModelForCausalLM.from_pretrained(
-    model_id,
-    device_map="auto",
-    torch_dtype=torch.float32,  # ✅ use float32 on CPU
+    model_id, torch_dtype=torch.float32, device_map="auto"
 )
 
-# Create text generation pipeline
+# Pipeline for text generation
 generator = pipeline(
     "text-generation",
     model=model,
     tokenizer=tokenizer,
-    do_sample=True,  # ✅ enables temperature to apply
+    do_sample=True,
     temperature=0.7,
 )
 
 
-# Chat handler
-def chat(message, history=[]):
-    prompt = f"<|user|>\n{message}\n<|assistant|>\n"
-    response = generator(prompt, max_new_tokens=256)
-    reply = response[0]["generated_text"].replace(prompt, "").strip()
-    return reply
+# Chat handler that returns OpenAI-style messages
+def chat(message, history):
+    prompt = ""
+    for turn in history:
+        prompt += f"<|user|>\n{turn['content']}\n<|assistant|>\n{turn['response']}\n"
+    prompt += f"<|user|>\n{message}\n<|assistant|>\n"
+    output = generator(prompt, max_new_tokens=256)[0]["generated_text"]
+    reply = output.replace(prompt, "").strip()
+    return {"response": reply}
 
 
-# Gradio Chat UI
+# Launch with OpenAI-style message format
 iface = gr.ChatInterface(
     fn=chat,
+    chatbot=gr.Chatbot(type="messages"),  # Matches interface default
     title="Phi-2 Chatbot",
-    chatbot=gr.Chatbot(type="messages"),  # ✅ future-proof Gradio
 )
 
 if __name__ == "__main__":

requirements.txt

@@ -1,4 +1,3 @@
-huggingface_hub==0.25.2
 transformers
 torch
 gradio