Create app.py

app.py

@@ -0,0 +1,40 @@
+import spaces
+
+import torch
+
+import gradio as gr
+
+from transformers import AutoTokenizer, AutoModelForCausalLM
+
+max_seq_length = 2048
+
+tokenizer = AutoTokenizer.from_pretrained("ua-l/gemma-2-9b-legal-steps200-merged-16bit-uk")
+model = AutoModelForCausalLM.from_pretrained(
+    "ua-l/gemma-2-9b-legal-steps200-merged-16bit-uk",
+    quantization_config=quantization_config,
+    device_map='auto'
+)
+
+
+@spaces.GPU
+def predict(question):
+    inputs = tokenizer(
+    [f'''### Question:
+    {question}
+    
+    ### Answer:
+'''], return_tensors = "pt").to("cuda")
+
+    outputs = model.generate(**inputs, max_new_tokens = 128)
+    
+    results = tokenizer.batch_decode(outputs, skip_special_tokens=True)
+
+    return results[0]
+
+inputs = gr.Textbox(lines=2, label="Enter a question", value="Як отримати виплати ВПО?")
+
+outputs = gr.Textbox(label="Answer")
+
+demo = gr.Interface(fn=predict, inputs=inputs, outputs=outputs)
+demo.launch()
+