README.md

metadata

license: gemma
datasets:
  - HuggingFaceH4/CodeAlpaca_20K
language:
  - en
base_model:
  - google/gemma-3-270m-it
pipeline_tag: text-generation
tags:
  - text-generation-inference
  - gemma
  - code
  - slm

GemCod270M - Aquamarine (gemma-270m-it-code v3.0.0)

GemCod is a lightweight code generation model finetuned using SFT on the base gemma-270m-it (https://huggingface.co/google/gemma-3-270m-it) model. It offers accurate and quick(ish) code snippet generation in all major programming languages. It's small size (270M parameters) allows it to run comfortably on laptop grade GPUs.

---

Estimated parameters: ~270M

Architecture: Gemma3

Intended use: Code snippet generation from natural language

---

Training data

• Source: CodeAlpaca_20K dataset (https://huggingface.co/datasets/HuggingFaceH4/CodeAlpaca_20K)
• Rows: ~20,000 rows templated with a custom .jinja chat format
• Training: trained for 4,000 steps on an RTX 3050 (4GB VRAM)

Usage

Install requirements:

pip install -r requirements.txt
pip install transformers datasets accelerate safetensors

Usage (Hugging Face Hub)

You can load it directly from HuggingFace:

from transformers import AutoTokenizer, AutoModelForCausalLM


tokenizer = AutoTokenizer.from_pretrained("DireDreadlord/GemCod-codegen-270M")
model = AutoModelForCausalLM.from_pretrained("DireDreadlord/GemCod-codegen-270M")
model.to(device)
model.eval()
model.resize_token_embeddings(len(tokenizer))


if tokenizer.pad_token is None:
    tokenizer.pad_token = tokenizer.eos_token

chat_template = """{% for message in messages %}{% if message['role'] == 'user' %}User: {{ message['content'] }}
{% elif message['role'] == 'assistant' %}Assistant: {{ message['content'] }}
{% endif %}{% endfor %}"""
tokenizer.chat_template = chat_template

def generate_code(prompt, max_tokens) -> str:
    messages = [
        {
            "role": "user",
            "content": prompt
        }
    ]
    
    formatted_prompt = tokenizer.apply_chat_template(
        messages,
        tokenize=False,
        add_generation_prompt=True 
    )

    
    inputs = tokenizer(formatted_prompt, return_tensors="pt").to(device)
    input_length = inputs["input_ids"].shape[1]
    
    with torch.no_grad():
        outputs = model.generate(
            **inputs,
            max_new_tokens=256,
            do_sample=False,
            num_beams=1,
            pad_token_id=tokenizer.eos_token_id,
            eos_token_id=tokenizer.eos_token_id,
            use_cache=False,
        )

    generated_tokens = outputs[0][input_length:]
    generated_text = tokenizer.decode(generated_tokens, skip_special_tokens=True)
    return generated_text


prompt = "give me a cpp function that prints the first n fibonacci numbers"
print("Prompt: ", prompt)

result = generate_code(prompt)
print(result)

Limitations

• Trained on curated dataset of 20,000 rows may cause mistakes in generation.
• Model for experimental use only; users should employ it as such under license.