Update app.py

app.py

@@ -1,7 +1,17 @@
 import gradio as gr
+import transformers
+import peft
 
-def greet(name):
-    return "Hello " + name + "!!"
+model = 'freQuensy23/toxic-llama2'
+model = peft.AutoPeftModelForCausalLM.from_pretrained(model_id)
+tokenizer = transformers.AutoTokenizer.from_pretrained(model_id)
 
-iface = gr.Interface(fn=greet, inputs="text", outputs="text")
+
+def generate(text, temp):
+    input_ids = tokenizer(f"User: {text}\nBot:", return_tensors='pt').input_ids
+    generated_ids = model.generate(input_ids=input_ids.to(model.device), temperature=temp, max_new_tokens=64)[0][len(input_ids[0]):]
+    return tokenizer.decode(generated_ids).split('\n')[0]
+
+iface = gr.Interface(concurrency_limit=2, fn=generate, inputs=[gr.Textbox(lines=5, placeholder="Type your prompt here...", value='''I am clever?'''), gr.Slider(0.1, 1.5, value=1.1)], 
+                     outputs=gr.Textbox())
 iface.launch()