LoRA and QLoRA concepts in Prompt Engineering / GenAI - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

What is the main purpose of LoRA in machine learning?

LoRA (Low-Rank Adaptation) is a technique used in fine-tuning large models. What is its main goal?

A To train the model from scratch without pre-trained weights

B To increase the size of the model by adding more layers

C To reduce the number of trainable parameters by adding low-rank matrices to existing weights

D To replace the entire model with a smaller one

💡 Hint

Think about how LoRA helps with training efficiency.

LoRA adds small low-rank matrices to existing model weights, allowing fine-tuning with fewer trainable parameters and less memory.

🧠 Conceptual

intermediate

How does QLoRA differ from LoRA?

QLoRA is an extension of LoRA. What key feature distinguishes QLoRA from standard LoRA?

A QLoRA uses quantized weights to reduce memory usage during fine-tuning

B QLoRA trains models without any pre-trained weights

C QLoRA removes low-rank matrices and trains full weights

D QLoRA increases the rank of adaptation matrices compared to LoRA

💡 Hint

Consider how QLoRA manages memory differently than LoRA.

QLoRA applies quantization to model weights, reducing memory use while still fine-tuning with low-rank adapters.

❓ Model Choice

advanced

Which model architecture is best suited for applying LoRA and QLoRA?

You want to fine-tune a large language model efficiently using LoRA or QLoRA. Which architecture is most appropriate?

A Decision trees

B Simple linear regression models

C Convolutional neural networks for image classification

D Transformer-based models with large dense layers

💡 Hint

LoRA and QLoRA are designed for large models with many parameters.

LoRA and QLoRA target large transformer models, which have many dense layers suitable for low-rank adaptation.

❓ Metrics

advanced

What metric best indicates successful fine-tuning with LoRA or QLoRA?

After fine-tuning a language model with LoRA or QLoRA, which metric best shows the model learned well without overfitting?

A Validation loss decreasing and stabilizing

B Training loss increasing steadily

C Validation accuracy dropping sharply

D Training accuracy staying constant at zero

💡 Hint

Good fine-tuning means the model improves on unseen data.

Validation loss decreasing and stabilizing means the model generalizes well after fine-tuning.

🔧 Debug

expert

Why might QLoRA fine-tuning fail with a 4-bit quantized model?

You try to fine-tune a large model using QLoRA with 4-bit quantization but get poor results. What is a likely cause?

A The model is too small to benefit from quantization

B Quantization noise is too high, degrading model performance during fine-tuning

C LoRA adapters are not compatible with quantized weights

D The optimizer does not support low-rank matrices

💡 Hint

Think about how very low-bit quantization affects model precision.

4-bit quantization can introduce noise that harms fine-tuning quality if not carefully managed.

Practice

1. What is the main purpose of LoRA in training large AI models?

easy

A. To increase the size of the model for better accuracy

B. To add small trainable parts that make training easier and cheaper

C. To replace the entire model with a smaller one

D. To remove layers from the model to speed up training

Solution

1. Step 1: Understand LoRA's role in model training

   LoRA adds small trainable parts to a big model instead of retraining the whole model, making training easier and cheaper.

2. Step 2: Compare options with LoRA's purpose

   Options B, C, and D describe changing model size or structure, which is not what LoRA does.

3. Final Answer:

   To add small trainable parts that make training easier and cheaper -> Option B

4. Quick Check:

   LoRA = small trainable parts for easier training [OK]

Hint: LoRA adds small parts to big models for easier training [OK]

Common Mistakes:

• Thinking LoRA replaces the whole model
• Confusing LoRA with model size increase
• Assuming LoRA removes layers

2. Which of the following correctly describes QLoRA?

easy

A. A method that combines LoRA with quantization to save memory

B. A technique that trains models without any compression

C. A way to increase model size by adding layers

D. A method that removes LoRA parts to speed up training

Solution

1. Step 1: Recall QLoRA's definition

   QLoRA combines LoRA with quantization (number compression) to reduce memory use and speed up training.

2. Step 2: Eliminate incorrect options

   Options B, C, and D contradict QLoRA's purpose by ignoring compression or removing LoRA parts.

3. Final Answer:

   A method that combines LoRA with quantization to save memory -> Option A

4. Quick Check:

   QLoRA = LoRA + quantization for memory saving [OK]

Hint: QLoRA = LoRA plus compression to save memory [OK]

Common Mistakes:

• Ignoring quantization in QLoRA
• Thinking QLoRA removes LoRA parts
• Believing QLoRA increases model size

3. Given this Python snippet using LoRA and QLoRA concepts:

model_size = 1000  # in MB
lora_size = 10    # LoRA adds 10 MB
quantization_factor = 0.25  # QLoRA compresses to 25%

lora_model_size = model_size + lora_size
qlora_model_size = int(lora_model_size * quantization_factor)
print(qlora_model_size)

What is the printed output?

medium

A. 252

B. 250

C. 260

D. 275

Solution

1. Step 1: Calculate LoRA model size

   LoRA adds 10 MB to 1000 MB, so lora_model_size = 1000 + 10 = 1010 MB.

2. Step 2: Apply QLoRA compression

   QLoRA compresses to 25%, so qlora_model_size = int(1010 * 0.25) = int(252.5) = 252 MB.

3. Final Answer:

   252 -> Option A

4. Quick Check:

   1010 * 0.25 = 252.5 -> 252 [OK]

Hint: Add LoRA size, then multiply by compression factor [OK]

Common Mistakes:

• Multiplying before adding LoRA size
• Rounding incorrectly
• Using 0.2 instead of 0.25 for compression

4. This code tries to calculate QLoRA model size but has an error:

model_size = 800
lora_size = 20
quantization_factor = 0.3

qlora_model_size = model_size + lora_size * quantization_factor
print(qlora_model_size)

What is the error and how to fix it?

medium

A. Wrong variable name; change quantization_factor to quant_factor

B. No error; code is correct

C. Should use integer division // instead of *

D. Missing parentheses; fix with (model_size + lora_size) * quantization_factor

Solution

1. Step 1: Identify operator precedence issue

   Multiplication (*) happens before addition (+), so only lora_size is multiplied by quantization_factor, not the sum.

2. Step 2: Fix with parentheses

   Use (model_size + lora_size) * quantization_factor to multiply the total size by compression factor.

3. Final Answer:

   Missing parentheses; fix with (model_size + lora_size) * quantization_factor -> Option D

4. Quick Check:

   Parentheses fix operator order [OK]

Hint: Use parentheses to control addition before multiplication [OK]

Common Mistakes:

• Ignoring operator precedence
• Changing variable names incorrectly
• Using wrong operators like //

5. You want to fine-tune a large language model on a small laptop with limited memory. Which approach best balances training speed and memory use?

hard

A. Only add LoRA layers without any compression

B. Train the full large model from scratch without compression

C. Use QLoRA to compress the model and add LoRA layers for efficient training

D. Use full precision training without LoRA or compression

Solution

1. Step 1: Understand resource limits

   Small laptops have limited memory, so full model training or full precision is too heavy.

2. Step 2: Choose best method

   QLoRA combines LoRA's small trainable parts with quantization compression, saving memory and speeding training.

3. Step 3: Compare options

   Options B and D ignore memory limits; A lacks compression benefits.

4. Final Answer:

   Use QLoRA to compress the model and add LoRA layers for efficient training -> Option C

5. Quick Check:

   QLoRA = LoRA + compression for small devices [OK]

Hint: Combine LoRA and compression for small device training [OK]

Common Mistakes:

• Ignoring compression benefits
• Trying full model training on small memory
• Using only LoRA without compression