Self-hosted LLMs (Llama, Mistral) in Prompt Engineering / GenAI - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Understanding Model Size Impact on Self-hosted LLMs

Which of the following best explains how increasing the number of parameters in a self-hosted LLM like Llama or Mistral affects its performance and resource requirements?

A Increasing parameters reduces accuracy but speeds up inference and lowers memory use.

B Increasing parameters decreases both accuracy and resource requirements.

C Increasing parameters has no effect on accuracy but increases training time only.

D Increasing parameters improves model accuracy but requires more memory and slower inference.

💡 Hint

Think about how bigger models usually behave in terms of accuracy and hardware needs.

Larger models generally capture more complex patterns, improving accuracy, but they need more memory and compute, slowing down inference.

❓ Predict Output

intermediate

Output of Token Generation with Temperature in LLM

Given the following pseudocode for generating tokens from a self-hosted LLM with temperature=0.0, what is the expected behavior of the output tokens?

Prompt Engineering / GenAI

tokens = model.generate(input_ids, temperature=0.0, max_length=5)
print(tokens)

A The model outputs tokens only from a fixed vocabulary subset.

B The model outputs the most likely tokens deterministically.

C The model outputs random tokens with equal probability.

D The model outputs tokens with high randomness and diversity.

💡 Hint

Temperature controls randomness in token selection.

A temperature of 0.0 means the model picks the highest probability token every time, making output deterministic.

❓ Model Choice

advanced

Choosing a Self-hosted LLM for Low-latency Applications

You want to deploy a self-hosted LLM for a chatbot that requires fast responses on limited hardware. Which model choice is best?

A A smaller Mistral 7B model quantized to 4-bit precision.

B A large Llama 70B model with full precision weights.

C A large Mistral 30B model with no quantization.

D A medium Llama 13B model with float32 precision.

💡 Hint

Consider model size and quantization effects on speed and memory.

Smaller models with quantization reduce memory and speed up inference, suitable for low-latency on limited hardware.

❓ Metrics

advanced

Evaluating Self-hosted LLM Output Quality

Which metric is most appropriate to evaluate the quality of text generated by a self-hosted LLM like Llama or Mistral on a language generation task?

A Perplexity measuring how well the model predicts the next token.

B Accuracy measuring exact token matches with ground truth.

C Mean Squared Error between predicted and actual token embeddings.

D F1 score measuring classification correctness.

💡 Hint

Think about metrics used in language modeling tasks.

Perplexity measures how surprised the model is by the next token, indicating prediction quality in language generation.

🔧 Debug

expert

Debugging Memory Error in Self-hosted LLM Inference

You try to run inference on a Llama 13B model but get a CUDA out-of-memory error. Which action will most likely fix this issue?

A Add more layers to the model to distribute memory load.

B Increase learning rate to speed up training and reduce memory.

C Reduce batch size or use model quantization to lower memory use.

D Disable GPU and run inference on CPU only.

💡 Hint

Think about how to reduce GPU memory usage during inference.

Reducing batch size or quantizing the model reduces memory needed, preventing out-of-memory errors during inference.

Practice

1. What is the main advantage of using self-hosted LLMs like Llama or Mistral?

easy

A. You keep full control and privacy over your data

B. They always run faster than cloud models

C. They require no installation or setup

D. They provide unlimited free internet access

Solution

1. Step 1: Understand self-hosted LLMs purpose

   Self-hosted LLMs run on your own machines, so your data stays private and under your control.

2. Step 2: Compare options

   Cloud models may send data externally; self-hosted models avoid this, ensuring privacy.

3. Final Answer:

   You keep full control and privacy over your data -> Option A

4. Quick Check:

   Privacy and control = B [OK]

Hint: Self-hosted means data stays with you, so privacy is key [OK]

Common Mistakes:

• Thinking self-hosted models are always faster
• Assuming no setup is needed
• Confusing self-hosted with cloud services

2. Which Python code snippet correctly loads a Llama model using the Hugging Face Transformers library?

easy

A. from transformers import LlamaForCausalLM; model = LlamaForCausalLM.from_pretrained('llama-model')

B. import llama; model = llama.load('llama-model')

C. from transformers import MistralModel; model = MistralModel.load('llama-model')

D. model = load_model('llama-model')

Solution

1. Step 1: Identify correct library and class

   The Hugging Face Transformers library uses LlamaForCausalLM to load Llama models.

2. Step 2: Check method to load model

   from_pretrained is the standard method to load pretrained models in Transformers.

3. Final Answer:

   from transformers import LlamaForCausalLM; model = LlamaForCausalLM.from_pretrained('llama-model') -> Option A

4. Quick Check:

   Transformers + from_pretrained = C [OK]

Hint: Use Transformers library and from_pretrained to load models [OK]

Common Mistakes:

• Using wrong import names
• Calling non-existent load methods
• Confusing Mistral and Llama classes

3. Given this code snippet using a Mistral model, what will be the output type of output?

from transformers import MistralForCausalLM, MistralTokenizer
model = MistralForCausalLM.from_pretrained('mistral-base')
tokenizer = MistralTokenizer.from_pretrained('mistral-base')
inputs = tokenizer('Hello world', return_tensors='pt')
outputs = model.generate(**inputs)
output = tokenizer.decode(outputs[0])

medium

A. An error because generate is not defined

B. A tensor of token IDs

C. A list of token probabilities

D. A decoded string of generated text

Solution

1. Step 1: Understand model.generate output

   model.generate returns token IDs as tensors representing generated text tokens.

2. Step 2: Decode tokens to string

   tokenizer.decode converts token IDs to a readable string.

3. Final Answer:

   A decoded string of generated text -> Option D

4. Quick Check:

   generate + decode = string output [OK]

Hint: generate returns tokens; decode converts tokens to string [OK]

Common Mistakes:

• Thinking output is raw tensor
• Confusing probabilities with tokens
• Assuming generate method is missing

4. You try to load a Llama model with this code but get an error:

from transformers import LlamaForCausalLM
model = LlamaForCausalLM.load('llama-model')

What is the likely cause of the error?

medium

A. LlamaForCausalLM cannot be imported from transformers

B. The model name 'llama-model' is invalid

C. The method load() does not exist; should use from_pretrained()

D. You need to install the Mistral library first

Solution

1. Step 1: Check method names in Transformers

   Transformers models use from_pretrained() to load models, not load().

2. Step 2: Identify error cause

   Using load() causes AttributeError because it is not defined for LlamaForCausalLM.

3. Final Answer:

   The method load() does not exist; should use from_pretrained() -> Option C

4. Quick Check:

   Use from_pretrained, not load [OK]

Hint: Use from_pretrained() to load models, not load() [OK]

Common Mistakes:

• Assuming load() is valid method
• Blaming model name without checking method
• Confusing Llama and Mistral imports

5. You want to run a self-hosted Llama model on your local machine but it has limited RAM. Which approach helps reduce memory usage while keeping reasonable performance?

hard

A. Use a cloud service instead of local hosting

B. Use quantization to reduce model size and load with 8-bit precision

C. Run the model on CPU without any batching

D. Load the full 32-bit model without any optimization

Solution

1. Step 1: Understand memory constraints

   Limited RAM means loading full 32-bit models is heavy and slow.

2. Step 2: Apply quantization

   Quantization reduces model size by using lower precision (e.g., 8-bit), saving memory and keeping decent speed.

3. Step 3: Evaluate other options

   Loading full model wastes memory; CPU without batching is slow; cloud is not self-hosted.

4. Final Answer:

   Use quantization to reduce model size and load with 8-bit precision -> Option B

5. Quick Check:

   Quantization saves memory and keeps performance [OK]

Hint: Quantize models to 8-bit for less RAM use [OK]

Common Mistakes:

• Loading full 32-bit model ignoring RAM limits
• Running without batching causing slow speed
• Switching to cloud defeats self-hosting purpose