Update app.py

app.py

@@ -119,7 +119,7 @@ def generate_recipe(user_inputs):
         text = f"{prompt}\n\nTools:\n{tool_section}"
         
         # Tokenize and move to the correct device
-        model_inputs = tokenizer([text], return_tensors="pt")
+        model_inputs = tokenizer(prompt, return_tensors="pt")
         # text = tokenizer.apply_chat_template(
         #     messages,
         #     tokenize=False,
@@ -132,15 +132,15 @@ def generate_recipe(user_inputs):
         torch.cuda.empty_cache()
         with torch.no_grad():
             generated_ids = model.generate(
-                model_inputs.input_ids,
+                **model_inputs,
                 max_new_tokens=512,
             )
 
-        generated_ids = [
-            output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
-        ]
+        # generated_ids = [
+        #     output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
+        # ]
 
-        st.session_state.recipe = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
+        st.session_state.recipe = tokenizer.decode(generated_ids[0], skip_special_tokens=True)
         st.session_state.recipe_saved = False
         
 def clear_inputs():