pubertcs/Ornith-1.0-9B-IL2CPP-Decompiler

# Ornith-1.0-9B-IL2CPP-Decompiler

This model is a specialized distillation of Ornith-1.0-9B, fine-tuned specifically for IL2CPP decompilation and conversion tasks.

Training Details

• Base Model: Ornith-1.0-9B
• Dataset: 5,000 examples of IL2CPP conversion data generated by Deepseek-V4-Flash
• Method: LoRA fine-tuning (Rank 16, 2 epochs) followed by full weight merging
• Purpose: To create a highly specialized model for converting and decompiling IL2CPP code structures with high accuracy and consistency.

Usage

This model is optimized for IL2CPP tasks. For general coding or other tasks, please use the base Ornith-1.0-9B model.

from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "pubertcs/Ornith-1.0-9B-IL2CPP-Decompiler"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name, device_map="auto")

messages = [{"role": "user", "content": "Convert this C# script to IL2CPP..."}]
text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer(text, return_tensors="pt").to(model.device)
generated = model.generate(**inputs, max_new_tokens=512, temperature=0.6)
print(tokenizer.decode(generated[0], skip_special_tokens=True))

GGUF Quants

Quantized versions of this model are available at pubertcs/Ornith-1.0-9B-IL2CPP-Decompiler-GGUF.

Citation

@misc{ornith_9b,
    title = {{Ornith-1.0-9B}: Agentic Coding, Open to All},
    url = {https://deep-reinforce.com/ornith_1_0.html},
    author = {{DeepReinforce Team}},
    year = {2026}
}