--- license: apache-2.0 language: - en tags: - renpy - visual-novel - storytelling - creative-writing - llama-3 - qlora - finetuned - text-generation - natural-language-to-script - instruction-free inference: false widget: - text: "A detective wakes up in a town where no one remembers him." - text: "Every time she falls asleep, she wakes up in another version of her life." - text: "A cursed mirror swaps your life with your reflection." datasets: - custom model-index: - name: Secunda-0.5-RAW results: [] --- ```ascii ▄▄▄▄▄ ▄███▄ ▄█▄ ▄ ▄ ██▄ ██ ████ ▄████▄ █ ▀▄ █▀ ▀ █▀ ▀▄ █ █ █ █ █ █ █ █ █ ▄ ▀▀▀▀▄ ██▄▄ █ ▀ █ █ ██ █ █ █ █▄▄█ █ █ █▄▄▄▄▄ ▀▀▄▄▄▄▀ █▄ ▄▀ █▄ ▄▀ █ █ █ █ █ █ █ █ █ █ █ █ ▀███▀ ▀███▀ █▄ ▄█ █ █ █ ███▀ █ ████▀ ▐█ ▄▄▄▄█ ▀▀▀ █ ██ █ ▀ ⋆⋆୨୧˚ THE PRIMÉTOILE ENGINE ˚୨୧⋆｡˚⋆ — Visual Novel generation under starlight — ``` | Version | Type | Strengths | Weaknesses | Recommended Use | |-------------------------------------------------------|-----------------|---------------------------------------------------------------------------|----------------------------------------------------------------------|-----------------------------| | [Secunda-0.1-GGUF](https://huggingface.co/Yaroster/Secunda-0.1-GGUF) / [RAW](https://huggingface.co/Yaroster/Secunda-0.1-RAW) | Instruction | - Most precise
- Coherent code
- Perfected Modelfile | - Smaller context / limited flexibility | **Production / Baseline** | | [Secunda-0.3-F16-QA](https://huggingface.co/Yaroster/Secunda-0.3-F16-QA) | QA-based Input | - Acceptable for question-based generation | - Less accurate than 0.1
- Not as coherent | Prototyping (QA mode) | | [Secunda-0.3-F16-TEXT](https://huggingface.co/Yaroster/Secunda-0.3-F16-TEXT) | Text-to-text | - Flexible for freeform tasks | - Slightly off
- Modelfile-dependent | Experimental / Text rewrite | | [Secunda-0.3-GGUF](https://huggingface.co/Yaroster/Secunda-0.3-GGUF) | GGUF build | - Portable GGUF of 0.3 | - Inherits 0.3 weaknesses | Lightweight local testing | | [Secunda-0.5-RAW](https://huggingface.co/Yaroster/Secunda-0.5-RAW) | QA Natural | - Best QA understanding
- Long-form generation potential | - Inconsistent output length
- Some instability | Research / Testing LoRA | | [Secunda-0.5-GGUF](https://huggingface.co/Yaroster/Secunda-0.5-GGUF) | GGUF build | - Portable, inference-ready version of 0.5 | - Shares issues of 0.5 | Offline experimentation | | [Secunda-0.1-RAW](https://huggingface.co/Yaroster/Secunda-0.1-RAW) | Instruction | - Same base as 0.1-GGUF | - Same as 0.1 | Production backup | # ⋆⁺₊⋆Secunda ☾ 0.5 RAW⋆⁺₊⋆ # For more concise and controlled results, I recommend using [Secunda-0.1-GGUF](https://huggingface.co/Yaroster/Secunda-0.1-GGUF) (instruct) or [Secunda-0.3-GGUF](https://huggingface.co/Yaroster/Secunda-0.3-GGUF) (natural) **Secunda-0.5-RAW** is the third iteration of the Secunda visual novel generation series, built to transform **natural language story prompts** into fully structured, high-quality Ren'Py scripts — without relying on explicit instructions or formatting. It was fine-tuned on a carefully curated dataset of fictional `.rpy` scenes, each paired with its original narrative concept. This model is much heavier than its siblings: [Secunda-0.1-RAW](https://huggingface.co/Yaroster/Secunda-0.1-RAW), [Secunda-0.3-F16-QA](https://huggingface.co/Yaroster/Secunda-0.3-F16-QA) and [Secunda-0.3-F16-TEXT](https://huggingface.co/Yaroster/Secunda-0.3-F16-TEXT).
Secunda is now based on both direct QA and structured prompting. --- ## 🚀 Model Highlights * **Instruction-Free Generation**: Unlike Secunda-0.1, this version relies only on a natural prompt such as *"A detective wakes up in a town where no one remembers him."* * **Finetuned with QLoRA (FP16)**: Efficient low-resource finetuning using 16-bit precision on LLaMA 3.1 8B. * **Script-Style Outputs**: Generates structured `.rpy` files, often including characters, backgrounds, dialogue, and a `return`. * **More Creative Freedom**: Produces more diverse narratives and styles, occasionally multi-scene outputs. --- /!\ NO HUMAN-MADE DATA WAS USED TO TRAIN THIS AI ! Secunda takes much pride in making sure the training data is scripted ! /!\ If you like Visual Novels, please visit [itch.io](itch.io) and support independant creators ! ## 🧠 Training Details * **Base model**: `meta-llama/Meta-Llama-3.1-8B` * **Finetuning**: QLoRA 4-bit adapters (FP16 runtime) * **Data**: 800+ handcrafted prompt/script pairs in `.jsonl`, each with a raw natural language idea and a full `.rpy` script. * **Hardware**: NVIDIA RTX 4070 * **Training time**: \~12 hours, 20 epochs --- ## 📝 Prompt Style This model works **without instructions**. Just write your idea naturally: > "An abandoned café reopens every full moon to serve ghosts their last coffee." --- ## ⚠️ Known Issues * May occasionally produce multiple unrelated scripts in one output if token limit is too high. * Less deterministic than instruction-based versions — retries may help. * Not always guaranteed to end with `return` (you may add it manually). --- ## 📌 Tips * Keep prompts between 10–30 words for best coherence. * If the output includes multiple scenes, you can split them manually. * If needed, combine with `sanitize_output()` logic from Secunda-0.1 to postprocess outputs. --- ## 🔒 Privacy This model is private and intended for research and internal tooling for the **Primétoile** visual novel engine. --- ## 📚 License Apache 2.0 — For research, testing, and development only. --- ## ✨ Credits Trained and maintained by [Yaroster](https://huggingface.co/Yaroster) for the **Secunda** engine inside the Primétoile framework.