README.md

---
license: cc-by-nc-nd-4.0
pipeline_tag: text-generation
library_name: transformers
tags:
- python
- coder
- developer-tools
- programming
- llm
---

# FastBit-450M-DeepCoder

FastBit-450M-DeepCoder is a lightweight LLM designed for Python code generation and logic processing. this model is optimized for high-speed inference on low-resource hardware like the **Intel i3-4150**.

## ⚖️ Terms of Use (License)
This model is licensed under the **Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International (CC-BY-NC-ND 4.0)**.

* **Attribution**: You must give credit to the project.
* **Non-Commercial**: You may not use this model for commercial purposes.
* **No Derivatives**: **You are strictly prohibited from modifying, remixing, or fine-tuning these weights.**

## 🚀 Implementation
To run this model locally, use the following Python script. Note: This model uses a custom weight file named `nanorons.safetensors`.

```python
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "imsuprtwo2/FastBit-450M-DeepCoder"

tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    dtype=torch.float32,
    low_cpu_mem_usage=True,
    trust_remote_code=True
)

prompt = "def calculate_factorial(n):"
inputs = tokenizer(prompt, return_tensors="pt")

outputs = model.generate(**inputs, max_new_tokens=100)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))
\```

## 🛠 Project Details
* **Model Name**: FastBit-450M
* **Parameters**: 450 Million
* **Optimization**: DeepCoder architecture for Python-specific tasks.
* **Status**: Active development by MASA.