Create app.py

app.py

@@ -0,0 +1,55 @@
+import gradio as gr
+from transformers import AutoModelForCausalLM, AutoTokenizer
+import torch
+
+MODEL_NAME = "microsoft/DialoGPT-medium"
+
+print("🔄 Loading model from Hugging Face...")
+
+tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
+model = AutoModelForCausalLM.from_pretrained(MODEL_NAME)
+
+print("✅ Model loaded!")
+
+chat_history_ids = None
+
+def chat(user_input):
+    global chat_history_ids
+
+    new_input_ids = tokenizer.encode(
+        user_input + tokenizer.eos_token,
+        return_tensors='pt'
+    )
+
+    if chat_history_ids is not None:
+        bot_input_ids = torch.cat([chat_history_ids, new_input_ids], dim=-1)
+    else:
+        bot_input_ids = new_input_ids
+
+    chat_history_ids = model.generate(
+        bot_input_ids,
+        max_length=1000,
+        pad_token_id=tokenizer.eos_token_id,
+        do_sample=True,
+        top_k=50,
+        top_p=0.95,
+        temperature=0.75
+    )
+
+    response = tokenizer.decode(
+        chat_history_ids[:, bot_input_ids.shape[-1]:][0],
+        skip_special_tokens=True
+    )
+
+    return response
+
+
+iface = gr.Interface(
+    fn=chat,
+    inputs="text",
+    outputs="text",
+    title="Emalawi19 AI 🤖",
+    description="Chat with your AI powered by Hugging Face"
+)
+
+iface.launch()