Code generation agent design in Agentic AI - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to initialize the agent's model with a transformer architecture.

Agentic AI

agent_model = TransformerModel([1]=512, num_layers=6)

Drag options to blanks, or click blank then click option'

Alearning_rate

Bdropout_rate

Cbatch_size

Dhidden_size

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'learning_rate' instead of 'hidden_size' here.

Confusing 'batch_size' with model architecture parameters.

The 'hidden_size' parameter sets the size of the hidden layers in the transformer model.

2fill in blank

medium

Complete the code to tokenize input text before feeding it to the agent.

Agentic AI

tokens = tokenizer.[1](input_text)

Drag options to blanks, or click blank then click option'

Aencode

Bdecode

Cgenerate

Dtrain

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'decode' which converts tokens back to text.

Using 'generate' which is for producing output, not tokenizing.

Tokenization converts text into tokens using the 'encode' method.

3fill in blank

hard

Fix the error in the code to correctly generate code output from the agent.

Agentic AI

generated_code = agent.[1](input_tokens)

Drag options to blanks, or click blank then click option'

Atrain

Bgenerate_code

Cpredict

Dcompile

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'train' which is for learning, not generating output.

Using 'generate_code' which is not a standard method name.

The 'predict' method is used to get the agent's output given input tokens.

4fill in blank

hard

Fill both blanks to create a dictionary comprehension that maps function names to their code strings if the code length is less than 50.

Agentic AI

code_map = {func_name: [1] for func_name, [2] in functions.items() if len(code) < 50}

Drag options to blanks, or click blank then click option'

Acode

Bfunc_code

Ccode_str

Attempts:

3 left

💡 Hint

Common Mistakes

Using inconsistent variable names for the code string.

Confusing keys and values in the dictionary comprehension.

The comprehension maps function names to their code strings stored in 'code', iterating over 'functions.items()' with 'func_name' and 'code'.

5fill in blank

hard

Fill all three blanks to filter and map functions with names starting with 'gen' and code length over 30.

Agentic AI

filtered_funcs = { [1]: [2] for [3], code in all_functions.items() if [3].startswith('gen') and len(code) > 30 }

Drag options to blanks, or click blank then click option'

Aname

Bfunc_name

Ccode

Dfunc_code

Attempts:

3 left

💡 Hint

Common Mistakes

Using different variable names inconsistently.

Not matching the loop variables with the condition variables.

The comprehension uses 'func_name' as key and 'code' as value, iterating over 'func_name, code' in all_functions.items().