Update app.py

app.py

@@ -1,14 +1,15 @@
 import gradio as gr
-import spaces
-import torch
+from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
 
-zero = torch.Tensor([0]).cuda()
-print(zero.device) # <-- 'cpu' 🤔
+model_name = "tencent/Hunyuan-MT-7B"
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+model = AutoModelForSeq2SeqLM.from_pretrained(model_name, device_map="auto")
 
-@spaces.GPU
-def greet(n):
-    print(zero.device) # <-- 'cuda:0' 🤗
-    return f"Hello {zero + n} Tensor"
+def translate(text, src="vi", tgt="km"):
+    inputs = tokenizer(text, return_tensors="pt").to(model.device)
+    outputs = model.generate(**inputs, forced_bos_token_id=tokenizer.lang_code_to_id[tgt])
+    return tokenizer.decode(outputs[0], skip_special_tokens=True)
 
-demo = gr.Interface(fn=greet, inputs=gr.Number(), outputs=gr.Text())
+demo = gr.Interface(fn=translate, inputs="text", outputs="text",
+    title="Hunyuan-MT-7B Dịch Việt ↔ Khmer")
 demo.launch()