File size: 2,553 Bytes
b0fcef0 24236ce b0fcef0 24236ce | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 | ---
license: gemma
base_model: google/gemma-2-2b-it
tags:
- python
- dsa
- tutoring
- qlora
- peft
- gemma
---
# Python/DSA Tutor — Gemma-2-2B QLoRA
Fine-tuned version of google/gemma-2-2b-it for Python and DSA tutoring.
## Model Description
This model is fine-tuned using QLoRA (4-bit quantization + LoRA adapters) to function as a structured Python/DSA tutor. It follows a consistent Goal → Key Concept → Python Example → Checkpoint Question format.
## Training
- Base model: google/gemma-2-2b-it
- Method: QLoRA (SFT with TRL SFTTrainer)
- Dataset: 546 synthetic tutoring examples
- Categories: concept, problem_solving, debugging, confusion, misconception, refusal
- LoRA rank: 16, alpha: 32
- Learning rate: 5e-5
- Epochs: 3
## How to Load
```python
from transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig
from peft import PeftModel
import torch
bnb_config = BitsAndBytesConfig(
load_in_4bit=True,
bnb_4bit_quant_type="nf4",
bnb_4bit_compute_dtype=torch.bfloat16,
bnb_4bit_use_double_quant=True,
)
base_model = AutoModelForCausalLM.from_pretrained(
"google/gemma-2-2b-it",
quantization_config=bnb_config,
device_map="auto",
attn_implementation="eager",
)
tokenizer = AutoTokenizer.from_pretrained("Tharun241100/python-dsa-tutor-gemma2-2b-qlora")
model = PeftModel.from_pretrained(
base_model,
"Tharun241100/python-dsa-tutor-gemma2-2b-qlora"
)
model.eval()
```
## Inference
```python
def generate_response(model, tokenizer, prompt, max_new_tokens=400):
formatted = f"<start_of_turn>user\n{prompt}<end_of_turn>\n<start_of_turn>model\n"
inputs = tokenizer(formatted, return_tensors="pt").to(model.device)
with torch.no_grad():
outputs = model.generate(
**inputs,
max_new_tokens=max_new_tokens,
do_sample=True,
temperature=0.7,
top_p=0.9,
repetition_penalty=1.3,
)
return tokenizer.decode(outputs[0][inputs.input_ids.shape[-1]:], skip_special_tokens=True)
response = generate_response(model, tokenizer, "Explain how a stack works in Python")
print(response)
```
## Example Output
**Prompt**: Explain how a stack works in Python
**Response**:
**Goal**: You will learn what stacks are and why they are useful.
**Key Concept**: A stack is like a pile of plates where you can only add or remove from the top (LIFO - Last In First Out).
**Python Example**: [clean code example]
**Checkpoint Question**: Can you think of a real-world application that uses a stack?
|