Chapter 3
Setting up Llama.py
Step 1) Install Python 3.
Step 2) Install the requests library:
“pip install requests”
Step 3) In the terminal, request:
“ollama pull gemma3:4b”
Step 4) Open Notepad or another text editor (VS Code, etc.) and paste the prompt to the right.
Save a file named run_llama.py.
Step 5) Open Terminal (Mac) or Command Prompt / PowerShell (Windows). Go to the folder where you saved the file, and run:
“python run_llama.py”
Prompt
import time
import requests
MODEL = "gemma3:4b"
prompt = input("Enter your prompt:\n")
t0 = time.perf_counter()
r = requests.post(
"http://localhost:11434/api/generate",
json={
"model": MODEL,
"prompt": prompt,
"stream": False,
"options": {
"num_predict": 1024} }, )
t1 = time.perf_counter()
print("runtime_sec:", t1 - t0)
data = r.json()
# Token
print("prompt_eval_count:", data.get("prompt_eval_count"))
print("eval_count:", data.get("eval_count"))
# Time in nanoseconds
print("total_duration:", data.get("total_duration"))
print("prompt_eval_duration:", data.get("prompt_eval_duration"))
print("eval_duration:", data.get("eval_duration"))
print("\n--- Full response -\n")
print(data.get("response", ""))
Running Llama.py
Copy your prompt from your doc.
Paste it into the script when it asks for input.
Press Enter.
Copy and save these outputs into your data table:
runtime_sec
prompt_eval_count
eval_count
mark correct/incorrect separately after checking the answer key (MMS-12)
* Use the same laptop and same model tag for all trials. Close extra apps if possible.
Example Response
“response_len_chars: 272
prompt_eval_count: 61
eval_count: 160
total_duration: 22336796400
prompt_eval_duration: 1919932400
eval_duration: 19537118600
--- Full response ---
Explanation:
We use the product rule: d/dx [u(x)v(x)] = u'(x)v(x) + u(x)v'(x).
Let u(x) = x³ and v(x) = e^(2x). Then u'(x) = 3x² and v'(x) = 2e^(2x).
Therefore, f'(x) = 3x²e^(2x) + x³(2e^(2x)) = 3x²e^(2x) + 2x³e^(2x).
Final Answer: f'(x) = 3x²e^(2x) + 2x³e^(2x)”
Data Collection
Total_duration: Run time (sec)
Response_len_chars: Token_in
Prompt_eval_count: Token_out
Based on the explanation and the MMS-12, identify if your answer is correct.
-
Add a short summary or a list of helpful resources here.