Week 1: Lab 1 coding suggestions for prompts

Course Q&A Generative AI with Large Language Models
1

Hi everyone,
For the Lab 1 coding assignment, can someone offer/provide examples from a syntax point of view before running the cell and evaluating the results?
I understand the theory and objectives but my syntax is leading to errors.

Please suggest unrelated examples or ideas to help understand the syntax. Perhaps there is a related resource that can clarify everything?

Here is a location where this would come in handy:
“for i, index in enumerate(example_indices):\n”,
" dialogue = dataset[‘test’][index][‘dialogue’]\n",
" summary = dataset[‘test’][index][‘summary’]\n",
" \n",
" prompt = f"""\n",
“Dialogue:\n”,
“\n”,
“{dialogue}\n”,
“\n”,
“What was going on?\n”,
“"""\n”,
“\n”,

Thanks.

2 Likes
2

Hi Alain. Please provide what errors you’re referring to so the Mentors will have more context about the problem. Thanks!

1 Like
3

Hi Chris and friends,

I was able to avoid error messages by changing the wording within the following:

prompt = f"""

What time does the train come?

{dialogue}
Summary:

"""

I’m still trying to figure out examples of what could be included within a prompt even in a totally different example. I’m sure this would be helpful to most.

Thanks!

1 Like
4

Hi Alain,

We could try different prompts and get different results.
For e.g., in Section 3.2, you can change the code to below:

for i, index in enumerate(example_indices):
    dialogue = dataset['test'][index]['dialogue']
    summary = dataset['test'][index]['summary']
        
    prompt = f"""
Dialogue:

{dialogue}

What time does the train come?
"""

    inputs = tokenizer(prompt, return_tensors='pt')
    output = tokenizer.decode(
        model.generate(
            inputs["input_ids"], 
            max_new_tokens=50,
        )[0], 
        skip_special_tokens=True
    )

    print(dash_line)
    print('Example ', i + 1)
    print(dash_line)
    print(f'INPUT PROMPT:\n{prompt}')
    print(dash_line)
    print(f'BASELINE HUMAN SUMMARY:\n{summary}\n')
    print(dash_line)
    print(f'MODEL GENERATION - ZERO SHOT:\n{output}\n')

and the FLAN-T5 model tries its best to answer the prompt with the following:

 ---------------------------------------------------------------------------------------------------
 Example  1
 ---------------------------------------------------------------------------------------------------
 INPUT PROMPT:

 Dialogue:

 #Person1#: What time is it, Tom?
 #Person2#: Just a minute. It's ten to nine by my watch.
 #Person1#: Is it? I had no idea it was so late. I must be off now.
 #Person2#: What's the hurry?
 #Person1#: I must catch the nine-thirty train.
 #Person2#: You've plenty of time yet. The railway station is very close. It won't take more than twenty minutes to get there.

 What time does the train come?

 ---------------------------------------------------------------------------------------------------
 BASELINE HUMAN SUMMARY:
 #Person1# is in a hurry to catch a train. Tom tells #Person1# there is plenty of time.

 ---------------------------------------------------------------------------------------------------
 MODEL GENERATION - ZERO SHOT:
 The train comes at 9:30.

I think that helps?