Update app.py

app.py

@@ -21,36 +21,38 @@ def download_argos_model(from_code, to_code):
     argostranslate.package.install_from_path(package_to_install.download())
 
 def wingpt(model_name, sl, tl, input_text):
-        model = AutoModelForCausalLM.from_pretrained(
-           model_name,
-           torch_dtype="auto",
-           device_map="auto"
-        )
-        tokenizer = AutoTokenizer.from_pretrained(model_name)
-        
-        messages = [
-           {"role": "system", "content": f"Translate this from {sl} to {tl} language"}, 
-           {"role": "user", "content": input_text}
-        ]
-        
-        text = tokenizer.apply_chat_template(
-           messages,
-           tokenize=False,
-           add_generation_prompt=False
-        )
-        model_inputs = tokenizer([text], return_tensors="pt").to(model.device)
-        
-        generated_ids = model.generate(
-           **model_inputs,
-           max_new_tokens=512,
-           temperature=0
-        )
-        
-        generated_ids = [
-           output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
-        ]
-        
-        return tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
+    model = AutoModelForCausalLM.from_pretrained(
+       model_name,
+       torch_dtype="auto",
+       device_map="auto"
+    )
+    tokenizer = AutoTokenizer.from_pretrained(model_name)
+    
+    messages = [
+       {"role": "system", "content": f"Translate this from {sl} to {tl} language"}, 
+       {"role": "user", "content": input_text}
+    ]
+    
+    text = tokenizer.apply_chat_template(
+       messages,
+       tokenize=False,
+       add_generation_prompt=False
+    )
+    model_inputs = tokenizer([text], return_tensors="pt").to(model.device)
+    
+    generated_ids = model.generate(
+       **model_inputs,
+       max_new_tokens=512,
+       temperature=0
+    )
+    
+    generated_ids = [
+       output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
+    ]
+    print(tokenizer.batch_decode(generated_ids, skip_special_tokens=True))
+    result = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
+
+    return result.replace(input_text, '').strip()
 
 # App layout
 st.header("Text Machine Translation")