Maincoder-1B-ONNX is the ONNX-optimized version of Maincoder-1B, a code-focused language model optimized for code generation and completion tasks. This version enables fast inference using ONNX Runtime in Python and runs directly in the browser with Transformers.js.

Key Features

• ONNX Optimized: Efficient inference with ONNX Runtime and KV-cache support
• Cross-Platform: Run in Python, Node.js, or directly in the browser
• Code Generation: Optimized for Python code completion and generation tasks.
• Compact Size: 1 billion parameters, lightweight enough to run on consumer hardware.
• SOTA Performance: State-of-the-art performance on Python coding benchmarks HumanEval, HumanEval+ and MBPP+.

Benchmark Results

Model Overview

Maincoder uses a modern transformer decoder architecture with:

• Rotary Position Embeddings: With theta of 1,000,000.
• RMSNorm: Pre-normalization for stable training.
• Grouped Query Attention: 4:1 ratio of query to key-value heads.
• QK Normalization: RMSNorm applied to attention queries and keys.
• SwiGLU MLP: Gated linear units with SiLU activation.

Usage

Python (ONNX Runtime)

Installation

pip install optimum[onnxruntime] transformers

For GPU acceleration:

pip install optimum[onnxruntime-gpu]

Quick Start

from optimum.onnxruntime import ORTModelForCausalLM
from transformers import AutoTokenizer

# Load the ONNX model with KV-cache support
model = ORTModelForCausalLM.from_pretrained(
    "Maincode/Maincoder-1B-ONNX",
    file_name="decoder_with_past_model.onnx",
    use_cache=True
)

# Load the tokenizer
tokenizer = AutoTokenizer.from_pretrained("Maincode/Maincoder-1B-ONNX")

# Code completion example
prompt = '''def fibonacci(n: int) -> int:
    """Return the n-th Fibonacci number."""
'''

inputs = tokenizer(prompt, return_tensors="pt")
outputs = model.generate(
    **inputs,
    max_new_tokens=128,
    temperature=0.2,
    do_sample=True,
    pad_token_id=tokenizer.eos_token_id
)

print(tokenizer.decode(outputs[0], skip_special_tokens=True))

GPU Acceleration

from optimum.onnxruntime import ORTModelForCausalLM

model = ORTModelForCausalLM.from_pretrained(
    "Maincode/Maincoder-1B-ONNX",
    use_cache=True,
    file_name="decoder_with_past_model.onnx",
    provider="CUDAExecutionProvider"
)

---

JavaScript (Transformers.js)

Installation

npm install @huggingface/transformers

Node.js

import { AutoModelForCausalLM, AutoTokenizer } from '@huggingface/transformers';

// Load the tokenizer and model
const tokenizer = await AutoTokenizer.from_pretrained('Maincode/Maincoder-1B-ONNX');
const model = await AutoModelForCausalLM.from_pretrained('Maincode/Maincoder-1B-ONNX', {
    subfolder: '.',
    model_file_name: 'decoder_with_past_model',
    use_external_data_format: true,

});

// Code completion example
const prompt = `def fibonacci(n: int) -> int:
    """Return the n-th Fibonacci number."""
`;

const inputs = await tokenizer(prompt, { return_tensors: 'pt' });

const outputs = await model.generate({
    input_ids: inputs.input_ids,
    attention_mask: inputs.attention_mask,
    max_new_tokens: 128,
    temperature: 0.2,
    do_sample: true,
});

const decoded = tokenizer.decode(outputs[0], { skip_special_tokens: true });
console.log(decoded);

---

Code Completion Examples

# Function completion
prompt = '''def quicksort(arr: list) -> list:
    """Sort a list using the quicksort algorithm."""
'''

# Class completion
prompt = '''class BinarySearchTree:
    """A binary search tree implementation."""
    
    def __init__(self):
'''

# Algorithm implementation
prompt = '''def dijkstra(graph: dict, start: str, end: str) -> tuple:
    """Find the shortest path using Dijkstra's algorithm.
    
    Args:
        graph: Adjacency list representation of the graph
        start: Starting node
        end: Target node
    
    Returns:
        Tuple of (distance, path)
    """
'''

Additional Notes

Limitations

• Context length limited to 2,048 tokens
• Primarily optimized for Python, performance may vary on other languages
• May generate code with bugs or security issues - always review generated code
• Browser performance depends on device capabilities

Disclaimer: This model has not undergone any alignment or safety tuning (e.g., RLHF/RLAIF, DPO, or safety fine-tuning). Outputs may be unsafe or biased. Please use appropriate safeguards and evaluate carefully for your use case.

License

This model is released under the Apache 2.0 License.

Citation

@misc{maincoder2025,
  title        = {Maincoder-1B: A High-Performance 1B Parameter Coding Model},
  author       = {Maincode Team},
  year         = {2025},
  organization = {Maincode},
  howpublished = {\url{https://huggingface.co/Maincode/Maincoder-1B}}
}

Related Models

• Maincode/Maincoder-1B - Original PyTorch model

Contact

For questions, issues, or collaboration inquiries, please visit Maincode.