Create app.py

app.py

@@ -0,0 +1,19 @@
+
+import gradio as gr
+from transformers import pipeline, AutoTokenizer, AutoModelForCausalLM
+
+model_id = "meta-llama/Llama-3.1-8B-Instruct"
+
+tokenizer = AutoTokenizer.from_pretrained(model_id, use_auth_token=True)
+model = AutoModelForCausalLM.from_pretrained(model_id, device_map="auto", load_in_4bit=True, use_auth_token=True)
+
+pipe = pipeline("text-generation", model=model, tokenizer=tokenizer)
+
+def ai_assistant(command):
+    prompt = f"<|begin_of_text|><|start_header_id|>user<|end_header_id|>\n{command}<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n"
+    result = pipe(prompt, max_new_tokens=100)[0]["generated_text"]
+    return result.split("<|eot_id|>")[0].split("<|end_header_id|>\n")[-1].strip()
+
+demo = gr.Interface(fn=ai_assistant, inputs="text", outputs="text", title="Llama 3.1 AI Assistant", description="Ask your assistant to do anything")
+
+demo.launch()