35 7 months ago

This model is used to teach python tutorial at the "ai-core.hack-shak internal"

tools
ollama run comanderanch/Zekial-CodeTutor

Details

7 months ago

9c80d95d714a · 2.0GB ·

llama
·
3.21B
·
Q4_K_M
LLAMA 3.2 COMMUNITY LICENSE AGREEMENT Llama 3.2 Version Release Date: September 25, 2024 “Agreemen
**Llama 3.2** **Acceptable Use Policy** Meta is committed to promoting safe and fair use of its tool
You are CodeMentor, an expert Python programmer and an encouraging AI tutor. Primary task: begin a l
{ "stop": [ "<|start_header_id|>", "<|end_header_id|>", "<|eot_id|>"
<|start_header_id|>system<|end_header_id|> Cutting Knowledge Date: December 2023 {{ if .System }}{{

Readme

Zekial CodeTutor

Tag: comanderanch/zekial-codetutor:latest Base: comanderanch/Zekial-codemaster:latest Purpose: Beginner-friendly coding tutor that teaches step-by-step using Socratic prompts, tiny runnable examples, and short practice tasks.

Highlights

Beginner-first pedagogy: no prior knowledge assumed.

Socratic method: guides with questions, not just answers.

Micro-examples + practice: every concept has a small runnable snippet and a tiny exercise.

Concise + supportive tone: “Every master was once a beginner.”

Configurable: accepts typical sampling params (e.g., temperature, top_p). Pull

ollama pull comanderanch/zekial-codetutor:latest

Run (CLI)

ollama run comanderanch/zekial-codetutor:latest \
  'topic="variables"; language="python"; user_goal="print a name and age"; Start the first lesson.'

Run (curl, local API)

curl http://localhost:11434/api/generate -d '{
  "model": "comanderanch/zekial-codetutor:latest",
  "prompt": "topic=\"lists\"; language=\"python\"; user_goal=\"append items and print\"; Begin the lesson."
}'

Open WebUI

Select comanderanch/zekial-codetutor:latest as the model.

First message can include your lesson intent:

topic="for loops"; language="python"; user_goal="iterate and sum"; Start the first lesson.

Intended Behavior

The model is configured with a System prompt that enforces:

Method: Socratic questions → Simple analogies → One concept at a time → Tiny code → Small exercise.

Best practices:

Clarity Rule: write code for humans first

LEGO Brick Rule: small, reusable functions

Blueprint Rule: plan before coding

Detective Rule: test constantly

Tone: concise, polite, encouraging.

Catchphrase: “Every master was once a beginner.”

Prompting Patterns

Begin a new lesson

topic="<concept>"; language="<language>"; user_goal="<your immediate goal>"; Start the first lesson.

Ask for another example

Give one more tiny example of the same concept, then a 1-minute practice exercise.

Checkpoint before moving on

Wait for my confirmation before introducing the next concept.

Request stricter guidance

Use shorter steps, ask me a question first, then show a minimal code block.

Sampling Parameters (recommended defaults)

You can set these per request or in a Modelfile that FROM this model.

temperature: 0.2  (lower = more focused teaching)
top_p:       0.9

CLI example

ollama run -p 'topic="functions"; language="python"; user_goal="write a function add(a,b)"' \
  -t 0.2 -m comanderanch/zekial-codetutor:latest

Programmatic Usage

Python (requests)

import requests, json

payload = {
  "model": "comanderanch/zekial-codetutor:latest",
  "prompt": 'topic="dictionaries"; language="python"; user_goal="count word frequency"; Start the first lesson.'
}
r = requests.post("http://localhost:11434/api/generate", json=payload, stream=True)
for line in r.iter_lines():
    if line:
        print(json.loads(line.decode())["response"], end="")