improved description

app.py

@@ -27,7 +27,6 @@ with open("model_configs.json", "r") as f:
 
 # Extract instruction
 extract_input = model_config["extract_input"]
-
 terminators = [
     tokenizer.eos_token_id,
     tokenizer.convert_tokens_to_ids("<|eot_id|>"),
@@ -35,7 +34,7 @@ terminators = [
 
 
 @spaces.GPU
-def generate_instruction():
+def generate_instruction_response():
     instruction = pipeline(
         extract_input,
         max_new_tokens=2048,
@@ -45,11 +44,13 @@ def generate_instruction():
         top_p=1,
     )
 
-    return instruction[0]["generated_text"][len(extract_input) :].split("\n")[0]
+    sanitized_instruction = instruction[0]["generated_text"][
+        len(extract_input) :
+    ].split("\n")[0]
 
+    response_template = f"""<|begin_of_text|><|start_header_id|>user<|end_header_id|>\n\n{sanitized_instruction}<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n"""
 
-def generate_response(response_template):
-    return pipeline(
+    response = pipeline(
         response_template,
         max_new_tokens=2048,
         eos_token_id=terminators,
@@ -58,13 +59,7 @@ def generate_response(response_template):
         top_p=1,
     )
 
-
-def generate_instruction_response():
-    sanitized_instruction = generate_instruction()
-    response_template = f"""<|begin_of_text|><|start_header_id|>user<|end_header_id|>\n\n{sanitized_instruction}<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n"""
-
     user_message = sanitized_instruction
-    response = generate_response(response_template)
     assistant_response = response[0]["generated_text"][len(response_template) :]
 
     return user_message, assistant_response
@@ -72,7 +67,7 @@ def generate_instruction_response():
 
 title = "Magpie demo"
 description = """
-This Gradio demo allows you to explore the approach outlined in the Magpie paper. "Magpie is a data synthesis pipeline that generates high-quality alignment data. Magpie does not rely on prompt engineering or seed questions. Instead, it directly constructs instruction data by prompting aligned LLMs with a pre-query template for sampling instructions." Essentially, instead of prompting the model with a question or a starting query, this approach relies on the pre-query template of the model to generate instructions. Essentially, you are giving the model only the template up to the point where a user instruction would start, and then the model generates the instruction and the response. 
+This Gradio demo showcases the approach described in the Magpie paper. Magpie is a data synthesis pipeline that creates high-quality alignment data without relying on prompt engineering or seed questions. Instead, it generates instruction data by prompting aligned LLMs with a pre-query template. This method does not prompt the model with a question or starting query. Instead, it uses the model's pre-query template to generate instructions. Essentially, the model is given only the template until a user instruction starts, and then it generates the instruction and the response.
 
 In this demo, you can see how the model generates a user instruction and a model response.