app added.

app.py

@@ -0,0 +1,47 @@
+import torch
+import gradio as gr
+from transformers import pipeline, logging, AutoModelForCausalLM, AutoTokenizer
+
+model_name = "microsoft/phi-2"
+model = AutoModelForCausalLM.from_pretrained(
+    model_name,
+    trust_remote_code=True
+)
+model.config.use_cache = False
+
+tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
+tokenizer.pad_token = tokenizer.eos_token
+
+peft_model_folder = './ckpts'
+model.load_adapter(peft_model_folder)
+
+def generate_text(input_text, max_length):
+  pipe = pipeline(task="text-generation",model=model,tokenizer=tokenizer, max_length=max_length)
+  result = pipe(f"<s>[INST] {input_text} [/INST]")
+  return_answer = result[0]['generated_text']
+  return return_answer
+
+# Create a Gradio interface
+title = "Phi2-QLora."
+description = "A simple Gradio interface to demo Phi2 model finetuned on openassist dataset with Qlora."
+examples = [["What is Large Language Model?"],
+            ["Why Python is most popular Language?"],
+            ["How to do rice?"]]
+demo = gr.Interface(
+    generate_text,
+    inputs=[
+        gr.TextArea(label="Enter Question"),
+        gr.Slider(1, 200, value = 10, step=1, label="Max Length")
+    ],
+
+    outputs=[
+        gr.Textbox(label="Response from Phi2 Model: "),
+        gr.TextArea(label="Tokens")
+    ],
+    title=title,
+    description=description,
+    examples=examples,
+    cache_examples=False,
+    live=True
+)
+demo.launch()